Biogeography of Angraecum (Orchidaceae, Angraecinae), out of Madagascar Case-Dispersal

Université de Montréal

Systématique évolutive et Biogéographie de Angraecum (Orchidaceae, Angraecinae) à Madagascar

par Andriananjamanantsoa Herinandrianina Notahianjanahary

Centre sur la Biodiversité Institut de Recherche en Biologie Végétale Département de Sciences Biologiques Faculté des Arts et Sciences

Thèse présentée à la Faculté des Arts et Sciences en vue de l’obtention du grade de Philosophiæ Doctor en sciences biologiques

Novembre, 2015

Université de Montréal Faculté des études supérieures et postdoctorales

Ce mémoire intitulé:

Systématique évolutive et Biogéographie de Angraecum (Orchidaceae, Angraecinae) à Madagascar

Présenté par:

Andriananjamanantsoa Herinandrianina Notahianjanahary

a été évalué par un jury composé des personnes suivantes:

Luc Brouillet, directeur de recherche Edward Louis Jr., co-directeur de recherche Jeff Saarela, évaluateur externe François-Joseph Lapointe, membre du jury Simon Joly, président du jury

Résumé

Le genre Angraecum est un groupe d’orchidées tropicales qui compte environ 221 espèces réparties en Afrique subsaharienne, dans l’ouest de l’Océan Indien, et au Sri Lanka. Plus de la moitié des espèces se trouvent à Madagascar, dont au moins 90% sont endémiques à l’île. L’étude systématique et taxonomique du genre Angraecum a toujours été problématique à cause de sa grande diversité morphologique. Pour faciliter la classification, des sections ont été établies dont la plus connue est celle de Garay (1973), qui regroupe les espèces sous 19 sections. Plusieurs analyses phylogénétiques avaient montré que le genre Angraecum et les sections de Garay ne sont pas monophylétiques. Cependant, aucune révision systématique n’a été apportée à cause du faible échantillonnage dans ces analyses. En incorporant un plus grand nombre d'espèces et en ajoutant d’autres caractères morphologiques dans l’analyse, nous avons apporté une plus grande résolution à la reconstruction phylogénétique du groupe. Cette résolution concerne surtout les nœuds plus profonds qui représentent les différents clades à l’intérieur d'Angraecum, qui correspondent à des sections naturelles. A partir de ces clades, nous avons redéfini 14 sections monophylétiques toute en reconnaissant cinq nouvelles.

Grâce à cette nouvelle phylogénie d'Angraecum, nous avons pu étudier la diversification du genre et de la sous-tribu Angraecinae en utilisant des méthodes macroévolutives, notamment les roles joués par les traits floraux dans la spéciation, tout en l'interprétant grâce aux histoires géologique et paléoclimatique. Le modèle de diversification chez les Angraecinae semble avoir été celui communément rencontré dans les forêts tropicales humides, c’est-à-dire une diversification par accumulation graduelle d’espèces à travers le temps et non pas une radiation adaptative rapide, comme souvent observée chez des lignées animales malgaches. Plusieurs caractères morphologiques jouent un rôle important dans la diversification des espèces d'Angraecum. Le début de la diversification d'Angraecum à Madagascar coïncide avec le mouvement progressif de l’île vers le nord, l’établissement de la mousson dans la partie nord de l’île durant le Miocène, et l’expansion de la forêt tropicale malgache pendant cette période.

Notre étude de l’histoire biogéographique des Angraecinae suggère une origine malgache de la sous-tribu et du genre Angraecum. On observe de la dispersion à longue

distance à partir de Madagascar vers le reste du monde dans le genre Angraecum. La forêt tropicale humide du Nord Est de Madagascar est le point de départ de la diversification des espèces d'Angraecum. Le premier événement de dispersion a débuté à l’intérieur de l’île vers la fin du Miocène. Cet évènement est marqué par une migration du Nord Est vers le centre de Madagascar. Par ailleurs, la majorité des événements de dispersion à longue distance se sont produits durant le Pliocène-Pléistocène à partir soit du centre, soit du Nord Est de l'île. On assiste à des migrations indépendantes vers l’Afrique de l’est et les Comores d’une part, et vers les Mascareignes d’autre part. Un seul événement fondateur ayant conduit à l’apparition de la section Hadrangis est observé dans les Mascareignes. La saison cyclonique joue un rôle significatif dans la dispersion à longue distance des graines d’orchidées, comparée aux vents dominants qui soufflent dans la région ouest de l’Océan Indien, notamment l’alizé et la mousson. La similarité des niches écologiques a facilité l’expansion des espèces d'Angraecum dans les Comores et les Mascareignes.

Mots-clés: accumulation graduelle des lignées, Angraecum, Angraecinae, biogéographie, délimitation des sections, diversification, phylogénie, systématique, taxonomie

Abstract

Angraecum is a group of tropical orchids that includes ca. 221 species distributed in Sub-Saharan Africa, the western Indian Ocean region, and Sri Lanka. At least half of the species are found in Madagascar, 90% being endemic to the island. Taxonomic studies of Angraecum have always been problematic because of the great morphological diversity of the group. To facilitate classification, sections have been proposed, the best-known system to date being that of Garay (1973) that subdivides the genus into 19 sections. Previous phylogenetic studies had shown that genus Angraecum and Garay’s sections are not monophyletic. However, no systematic review was made because of a reduced species sample in these analyses. Using a greater sampling from Madagascar and adding morphological characters to the analyses, we brought greater resolution to the phylogenetic reconstruction of the group. This resolution mainly concerns the deeper nodes representing different clades within Angraecum, which basically correspond to natural sections. By using these clades, we redefined 14 sections and recognized five new ones.

Using this phylogeny of Angraecum, we evaluated species diversification using macroevolutionary methods, essentially the effect of floral traits in speciation. The great diversity of Angraecum species in Madagascar, the high endemicity, and the geology and paleoclimate histories allowed us to evaluate diversification patterns within the genus as well as sub-tribe Angraecinae. The model of diversification in Angraecinae follows that of most tropical rain forest taxa, which results from the gradual accumulation of species through time and not from a rapid adaptive radiation, as is often the case for Malagasy fauna lineages. Several morphological characters are involved in the diversification of Angraecum. The beginning of Angraecum diversification in Madagascar coincided with the progressive movement of the island northwards, the establishment of a monsoon regime in the northern part of the island during the Miocene, and the expansion of the Malagasy rainforest during that period.

Our historical biogeographic study of Angraecinae suggests a Malagasy origin of the subtribe Angraecinae and Angraecum. We observed out-of Madagascar long-distance dispersal in Angraecum. The north-eastern Malagasy rainforest is the center of species

iii

radiation for the genus. The first dispersal event within the island started in the late Miocene. This event was a migration from the north to the central highland. The majority of long- distance dispersal events outside Madagascar occurred during the Pliocene-Pleistocene, originating from either the center or the North East of the island. There were multiple independent dispersals to East Africa and the Comoros, and to the Mascarenes. A single founder-effect event in section Hadrangis is observed in the Mascarenes. The cyclonic seasons play a significant role in long-distance dispersal of orchid seeds, as compared to prevailing winds in the western Indian Ocean region, essentially trade wind and monsoon. Ecological niche similarity favored the expansion of Angraecum species in the Comoros and Mascarene archipelagos.

Keywords: accumulation of lineages through time, Angraecum, Angraecinae, biogeography, delimitation of sections, diversification, phylogeny, systematic, taxonomy

Table des matières

Résumé ...... i Abstract ...... iii Table des matières ...... v Liste des tableaux ...... ix Liste des figures ...... xi Liste des abréviations ...... xiv Remerciements ...... xvii Introduction ...... 1 1. Historique du genre Angraecum ...... 2 2. Classification et description d’Angraecum ...... 4 3. Biologie, écologie et répartition d’Angraecum ...... 5 4. Choix du site d’étude et échantillonnage ...... 6 5. Systématique et phylogénie d’Angraecum ...... 10 6. Diversification chez Angraecum ...... 11 7. Biogéographie historique d’Angraecum ...... 12 8. Problématiques, objectifs et hypothèses de recherche ...... 15 Chapitre 1: Diversification of Angraecum (Orchidaceae, Vandeae) in Madagascar: species accumulation through time rather than rapid radiation ...... 17 1.1. Résumé/ Abstract ...... 18 Résumé ...... 18 Abstract ...... 19 1.2. Introduction ...... 20 1.3. Material and methods ...... 22 1.3.1. Taxon sampling ...... 22 1.3.2. Morphological data ...... 23 1.3.3. PCR amplification and DNA sequencing ...... 23 1.3.4. Phylogenetic analyses ...... 24 1.3.5. Estimation of divergence times ...... 25

1.3.6. Diversification analyses ...... 26 1.3.7. Character state reconstruction ...... 28 1.4. Results ...... 28 1.4.1. Phylogenetic relationships ...... 28 1.4.2. Divergence time estimates ...... 34 1.4.3. Species diversification ...... 34 1.4.4. Ancestral state reconstructions ...... 39 1.5. Discussion ...... 41 1.5.1. Systematics of Angraecinae ...... 41 1.5.2. Systematics of Angraecum ...... 42 1.5.3. Labellum position a missing character to delineate sections in Angraecum ...... 44 1.5.4. Temporal framework and paleoclimate events in Angraecum ...... 44 1.5.5. Diversification in Angraecum ...... 45 1.6. Conclusion ...... 48 1.7. Acknowledgments ...... 49 Chapitre 2: Systematics and taxonomic revision of genus Angraecum (Orchidaceae, Angraecinae) and two new genera in the orchids of Madagascar ...... 50 2.1. Résumé/ Abstract ...... 51 Résumé ...... 51 Abstract ...... 51 2.2. Introduction ...... 52 2.3. Sections of Angraecum ...... 54 2.3.1. Angraecum Bory, Voy. 1: 359, t. 19 (1804)...... 54 2.3.2. Angraecum sect. Africanae Andriananjamanantsoa sect. nov...... 55 2.3.3. Angraecum sect. Angraecoides (Cordem.) Garay, Kew Bull. 28: 495–516 (1973)...... 56 2.3.4. Angraecum sect. Angraecum Benth., Benth. & Hook. F. Gen. Pl. 3: 583 (1883). . 56 2.3.5. Angraecum sect. Arachnangraecum Schltr., Fedde, Repert. Sp. Nov. Beih. 33: 309 (1925)...... 57 2.3.6. Angraecum sect. Boryangraecum Schltr., Fedde, Repert. Sp. Nov., Beih. 33: 308 (1925)...... 58

2.3.7. Angraecum sect. Hadrangis Schltr., Beih. Bot. Centralbl. 36(2): 158 (1918)...... 61 2.3.8. Angraecum sect. Humblotiangraecum Schltr., Fedde, Repert. Sp. Nov., Beih. 33: 310 (1925)...... 61 2.3.9. Angraecum sect. Nana (Cordem.) Garay, Kew Bull. 28: 495–516 (1973)...... 62 2.3.10. Angraecum sect. Oeoniella (Schltr.) Andriananjamanantsoa comb. nov...... 63 2.3.11. Angraecum sect. Perrierangraecum Schltr., Fedde, Repert. Sp. Nov., Beih. 33: 309 (1925)...... 64 2.3.12. Angraecum sect. Pseudojumellea Schltr., Beih. Bot. Centralbl. 36(2): 157 (1918)...... 66 2.3.13. Angraecum sect. Robusta Andriananjamanantsoa sect. nov...... 67 2.3.14. Angraecum sect. Sobennikoffia (Schltr.) Andriananjamanantsoa comb. nov...... 68 2.3.15. Angraecum sect. Stellariangraecum Andriananjamanantsoa sect. nov...... 69 2.3.16. Incertae sidis ...... 70 2.3.17. Species excluded from Angraecum ...... 70 2.4. New genera of Malagasy orchids ...... 71 2.4.1. Acaulia (Garay) Andriananjamanantsoa comb. nov...... 71 2.4.2. Parangraecum Andriananjamanantsoa gen. nov...... 72 2.5. Acknowledgements ...... 74 Chapitre 3: Biogeography of Angraecum (Orchidaceae, Angraecinae), out of Madagascar case-dispersal ...... 75 3.1. Résumé/Abstract ...... 76 Résumé ...... 76 Abstract ...... 77 3.2. Introduction ...... 78 3.3. Material and methods ...... 81 3.3.1. Phylogenetic inference and divergence time estimates ...... 81 3.3.2. Diversification analyses ...... 81 3.3.3. Biogeographical history analyses ...... 83 3.4. Results ...... 85 3.4.1. Divergence time estimates ...... 85 3.4.2. Habitat diversification ...... 87

vii

3.4.3. Ancestral range reconstruction ...... 88 3.5. Discussion ...... 94 3.5.1. Stratification changes and orchid diversification ...... 94 3.5.2. Malagasy origin of subtribe Angraecinae ...... 95 3.5.3. Origin of Angraecum and its dispersal within Madagascar ...... 96 3.5.4. Out of Madagascar dispersal of Angraecum ...... 97 3.5.5. Dispersal mechanisms and colonization in Angraecum ...... 99 3.6. Conclusion ...... 100 3.7. Acknowledgements ...... 101 Conclusion ...... 102 Délimitation du genre Angraecum sensu stricto ...... 102 Origine et diversité d’Angraecum à Madagascar ...... 103 Perspectives futures ...... 104 Bibliographie ...... i Annexe 1. Voucher and sources of DNA data ...... i Annexe 2. Morphological character description ...... ix Annexe 3. Morphological character matrix ...... xvi Annexe 4. Characters used in BAMM analyses ...... xxii Annexe 5. ITS2 phylogenetic tree ...... xxvii Annexe 6. cpDNA + morphology phylogenetic tree ...... xxx Annexe 7. Morphological phylogenetic tree ...... xxxiv Annexe 8. Dated phylogenetic tree ...... xxxvii Annexe 9. Spur length BAMM analysis ...... xli Annexe 10. Ancestral state reconstructions of floral traits ...... xlii Annexe 11. Speciation analyses ...... xlvi

viii

Liste des tableaux

Tableau 1. Liste et caractéristiques des sites d’étude...... 9 Tableau 2. Liste des 19 sections de Garay (1973) avec leur répartition géographique; les chiffres donnent le nombre d'espèces de chaque région et ceux entre paranthèses les espèces endémiques (Govaerts et al., 2015). Les sous-espèces et variétés ne sont pas incluses. Abréviations: Af: Afrique; E: est; C: centrale; O: ouest; Com: Comores; Mad: Madagascar; Msc: Mascareignes; Sey: Seychelles; Sri: Sri Lanka...... 10 Table 1.1. Comparaison of diversificaiton model used for the BiSSE and MuSSE analyses. Bold indicates the best fit model selected to test diversification of Angraecinae. Arrow represents transition allowed between states: ↔, reversible, →, irreversible. Parameters: λ, speciation rate; µ, extinction rate; q, character transition rate. Character states: A0, green; A1, white; B0, uppermost; B1, lowermost; C1, tiny; C2, small; C3, medium; C4, large; C5, very large; D1, very short; D2, short; D3, medium; D4, long; D5, very long. Abbreviations: AIC, Akaike information criterion; M, model tested; lnLik, Log likelihood...... 35 Table 3.1. Comparison of models tested in MuSSE; bold characters within the table highlight the best fit model selected for analyses. Character states: E, canopy epiphyte; I, inselberg lithophyte; L, undergrowth lithophyte; T, trunk epiphyte; ↔, reversible transition; →, irreversible transition. Abbreviations: AIC, Akaike information criterion; Df, degrees of freedom; lnLik, log likelihood; M, model...... 82 Table 3.2. Time slices and constraints used in ancestral area reconstructions of subtribe Angraecinae implemented in BioGeoBEARS; age of Mascarenes (2 Ma), Mayotte (7 Ma), and Comoros (15 Ma). Areas within Madagascar: MdN, north; MdNe, northeast; MdSe, southeast; MdC, center; MdW, west. Areas outside Madagascar: Ms, Mascarenes; Cm, Comoros; Se, Seychelles/Sri Lanka; AA, Africa/America/Asia. Abbreviations: M, model constraint; Ma, million years...... 84 Table 3.3. Ages of crown nodes of relevant clades presented in the phylogeny of Angraecinae inferred from BEAST (from chap. 1; Appendix 8). Comparison between the mean time to most recent common ancestor (MRCA) estimates and the node height highest posterior density (HPD) intervals at 95%...... 86

Table 3.4. Diversification rates of the forest strata character states obtained from the best transition model (5) selected in MuSSE (Table 3.1)...... 88 Table 3.5. Comparison of biogeographic models tested in BioGeoBEARS, bold represent selected model used for interpretation and discussions. Abbreviations: AIC, Akaike information criterion; AICwt, AIC weight; alt, alternative (model using parameter j); null, strandard model; LnL, log likelihood; J, parameter j including founder effect; M: constrained model with differential dispersal rate and time slices; pval, p-values...... 89

Liste des figures

Fig. 1. Comparaison entre Angraecinae et Aerangidinae. 1: colonne avec anthère; 2: rostellum; 3: pollinies (Dressler, 1993)...... 5 Fig. 2. Carte de végétation de Madagascar (Du Puy et Moat, 1996) ...... 7 Fig. 3. Diversification du genre Angraecum. (A) A. sedifolium, (B) A. danguyanum, (C) A. mirabile, (D) A. longicalcar, (E) A. sesquipedale (Photo: Andriananjamanantsoa H.)...... 12 Fig. 4. Position approximative de Madagascar dans le bloc du Gondwana-Est au Crétacé. Marges continentales interprétées à partir des données gravimétriques satellitaires (Eagles et Konig, 2008)...... 14 Fig. 1.1. Phylogenetic relationships within subtribe Angraecinae. 50% Bayesian majority-rule consensus tree from combined plastid data (matK, rps16 and trnL). Values above branches or at nodes represent posterior probability (PP) and bootstrap percentage (BP) support. Dashes represent branches that collapsed in MP strict consensus tree. Taxa with asterisk are Angraecum sensu Garay species. Abbreviations in brackets denote sections sensu Garay (1973): Aca = Acaulia, Agd = Angraecoides, Ang = Angraecum, Arc = Arachnangraecum, Bor = Boryangraecum, Chl = Chlorangraecum, Fil = Filangis, Gom = Gomphocentrum, Had = Hadrangis, Hum = Humblotiangraecum, Lem = Lemurangis, Lep = Lepervenchea, Pct = Pectinaria, Per = Perrierangraecum, Psj = Pseudojumellea...... 30 Fig. 1.2. Posterior probability distributions for the speciation rates (in Ma) of morphological characters using equal rate speciation (µ0 ~ µ1) with the BiSSE model and equal rate extinction (λi ~ λj) with the MuSSE model: flower color (A), labellum position (B), flower size (C), and spur length (D). Abbreviation: v, very...... 36 Fig. 1.3. Configuration shifts from the 95% credible set sampled by BAMM from the Angraecinae phylogeny and evolutionary rates through time. The intensity of colors on branches reflects the relative probability density of speciation rates (cool colors = slow, warm = fast). Black circles denote the position of the macroevolutionary regime shifts present in each sample. Blue curve indicates the mean speciation rate-through-time trajectory of Angraecinae in million years...... 37

Fig. 1.4. Best configuration shifts from the 95% credible set sampled by BAMM for the evolution of spur length across the phylogeny of Angraecinae. Color intensity on branches reflects the relative probability density of the instantaneous rate of phenotypic evolution. Black circles denote the position of the macroevolutionary regime shifts present in each sample. Blue curve denotes the mean evolution rate-through-time trajectory...... 38 Fig. 1.5. Ancestral state reconstructions of floral traits in Angraecinae implemented in ‘diversitree’; colors represent character states and pie charts represent the probability of ancestral states at nodes. Asterisks (*) indicate taxa illustrated in the pictures to the right to represent the flower shape of each clade except for Jumellea which is represented by Jumellea comorensis (not sampled in the phylogeny). Photo: Andriananjamanantsoa...... 39 Fig. 2.1. Partial phylogenetic relationships of Angraecum showing ancestral state reconstructions of labellum position implemented in ‘diversitree’ (chap. 1); complete tree presented in upper left corner, with selected partial trees in red-dashed boxes shown on the right; taxa in bold are the new genera Acaulia and Parangraecum. Colors represent labellum character states (orange, uppermost; green, lowermost) and pie charts represent the probability of ancestral states at node. Abbreviation: A, Angraecum...... 52 Fig. 2.2. Acaulia rhynchoglossa (Photo: Andriananjamanantsoa) ...... 72 Fig. 2.3. Parangraecum cf. humile (Photo: Andriananjamanantsoa) ...... 73 Fig. 3.1. Biogeographic areas inside Madagascar used in ancestral area reconstructions of subtribe Angraecinae implemented in BioGeoBEARS, colors indicate the different areas and letters denote the 22 administrative regions of Madagascar...... 85 Fig. 3.2. (A) Forest strata transition model selected in MuSSE. Arrows indicate direction of transitions; values next to arrows are transition rates; values in bold are diversification rates. E, canopy epiphytes; L, undergrowth lithophytes; T, trunk epiphytes; I, inselberg lithophytes. (B) Speciation rate of each character state...... 87 Fig. 3.3. Ancestral area reconstructions of subtribe Angraecinae from DECM1+J analysis implemented in BioGeoBEARS: A, reconstruction using eight geographic areas (more detailed outside Madagascar); B, reconstruction using nine geographic areas (more detailed within Madagascar); dash lines indicate time slices. Areas within Madagascar: MdC, center; MdN, north; MdNe, northeast; MdSe, southeast; MdW, west. Areas outside Madagascar: AA,

xii

Africa/America/Asia; Am, South America; As, Southeast Asia; eA, east Africa; wA, west Africa; Cm, Comoros; Ms, Mascarenes; Se, Seychelles/Sri Lanka...... 89 Fig. 3.4. Movements of dispersal in subtribe Angraecinae, arrows indicate the direction of the migration and numbers designate the age estimate of the events in million years. Red arrow indicates the early migration of subtribe Angraecinae from Southeast Asia towards Madagascar and Africa; blue arrows indicate the migration path of Angraecinae from Madagascar towards Africa then South America; green arrow denotes secondary migrations from Africa to Madagascar...... 96 Fig. 3.5. Out of Madagascar dispersal of Angraecum, arrows indicate the direction of the migration and numbers designate the age estimate of the events in million years...... 98

xiii

Liste des abréviations

AIC: Akaike information criterion alt: alternative APG: Angiosperm phylogeny group classification BA: Bayesian analyses BAMM: Bayesian analysis of macroevolutionary mixtures BEAST: Bayesian evolutionary analysis sampling trees BioGeoBEARS: biogeography with bayesian and likelihood evolutionary analysis in R scripts BiSSE: binary state speciation and extinction BS: bootstrap support BSA: bovine serum albumin

C3: plant metabolism using C3 carbon fixation CAM: crassulacean acid metabolism ChiSq: Chi-square CI: consistency index CITES: convention on international trade in endangered species of wild fauna and flora CTAB: cetyl trimethylammonium bromide DEC: dispersal-extinction-cladogenesis DIVA: dispersal–vicariance Df: degrees of freedom DNA: deoxyribonucleic acid e.g.: exempli gratia etc.: et cætera ExoSAP: exonuclease I and shrimp alkaline phosphatase HPD: highest probability densities IOI: Indian Ocean Islands ITS: internal transcribed spacer IUCN: international union for conservation of nature K: Kew herbarium lnLik: log likelihood

xiv

matK: chloroplast maturase K gene MCCT: maximum clade credibility tree MCMC: Markov chain Monte Carlo ML: maximum likelihood MP: maximum parsimony MRCA: most recent common ancestor MT: Marie-Victorin herbarium MuSSE: multistate speciation and extinction P: Paris herbarium PAUP*: phylogenetic analysis using parsimony *and other methods PBD: phylogenetic beta diversity PCR: polymerase chain reaction PP: posterior probability Pr: probability PVP: polyvinylpyrrolidone RHS: Royal horticultural society RI: retention index rps16: chloroplast ribosomal protein S16 s.l.: sensu lato s.s.: sensu stricto TAN: national herbarium of Madagascar TBR: tree bisection–reconnection trnL: chloroplast transfer ribonucleic acid asparagine-like WIOR: west Indian Ocean region wt: weight

Je dédie ce travail à:

ma femme, Herivonjy Ramaroson, pour son amour, sa patience et son encouragement tout au long de mes études,

ma sœur, Herizo N.F. Anjarasoa, pour son soutien

depuis toujours dans tous mes projets

xvi

Remerciements

Je tiens à remercier mon directeur de recherche, Dr. Luc Brouillet, et mon co-directeur, Dr. Edward Louis Jr., pour leur encadrement technique, leur soutien matériel et financier, et leur patience tout au long du projet.

Un grand merci à toute l’équipe du laboratoire de génétique du Center for Conservation and Research du Omaha’s Henry Doorly Zoo and Aquarium, en particulier Shannon Engberg, Runhua Lei, Cynthia Frasier, et Carolyn Bailey, pour leur grande contribution au séquencage de nos échantillons. Je remercie également le Canadian Centre for DNA Barcoding (CCDB) et le Barcode of Life Data Systems (BOLD), en particulier Masha Kuzmina, pour avoir contribué aussi au séquençage de nos échantillons.

Je tiens aussi à remercier toute l’équipe du Madagascar Biodiversity Partnership, agents de terrain, personnels de bureau, conducteurs, et étudiants gradués, pour leur soutien et collaboration dans le bon déroulement des travaux de terrain.

Je tiens à remercier le Département de biologie et écologie végétales de l’Université d’Antananarivo, le Madagascar National Park, et le Ministère de l'Environnement, de l'Écologie, de la Mer et des Forêts de Madagascar pour avoir facilité l’acquisition des permis de recherche.

Je remercie Dr. Anne Bruneau et Dr. Simon Joly pour leurs commentaires constructifs et leurs recommandations pour l’amélioration de mon projet. Je suis très reconnaissant du soutien technique apporté par Edeline Gagnon et Sébastien Renaut dans les analyses des données. Je remercie mes ami(e)s du Centre sur la biodiversité et de l’Institut de recherche en biologie végétale (IRBV), en particulier Rim Klabi, Aurélien Moreau, Marielle Babineau, Erin Zimmerman, Guillaume Bourdel, Boris Domenech, Gonzalo Bilbao, Hermine Alexandre, Yanitch Aymeric, et toute l’équipe du laboratoire de systématique qui étaient toujours là pour moi, surtout dans les moments les plus difficiles.

Je remercie également toutes les personnes qui ont contribué de près ou de loin à mon projet, Guy Suzon Ramangason, Eliane Ralambofetra, Bakolimalala Rakouth, Elisabeth Rabakonandrianina, Vololoniaina Jeannoda, Aro Vonjy Ramarosandratana, Vonjison

xvii

Rakotoarimanana, Hiariniaina Randrianizahana, Jacky Andriatiana, Nivo Raharsion, Hanitra Radamason, Stephane Bailleul, Denis Laperièrre, François Marquis, Geoffrey Hall, Kanchi Gandhi, Johan Hermans, Félix Forest, Claire Micheneau, Marc Chase, Tariq Stévart, Marc Pignal, André Schuiteman, Marilyn Light et Michael MacConaill pour leur soutien technique et recommandations.

Je remercie le Omaha’s Henry Doorly Zoo and Aquarium et le Madagascar Biodiversity Partnership, en particulier Dennis Pate et Edward Louis Jr., pour le soutien financier accordé à mes études, mes travaux de recherche, et mes déplacements. Je remercie également Les Amis du Jardin Botanique de Montréal, le Centre de la science de la biodiversité du Québec (CSBQ), la bourse Jacques Rousseau (IRBV), le Southern Ontario orchid society (SOOS), pour des bourses d’excellence et de voyage à l’étranger pour des conférences.

Je remercie ma femme, Herivonjy Ramaroson, pour son courage et sa patience durant toutes ces années de séparation. Je remercie également toute ma famille, mes parents, mon frère et mes sœurs, oncles et tantes, cousins et cousines, pour leur soutien financier et moral. Je suis reconnaissant envers mes familles d’accueil, ma tante et sa famille, Julie Marleau et sa famille, pour leur hospitalité.

xviii

Introduction

La famille des Orchidaceae, regroupant plus de 24000 espèces, est l’une des plus grandes familles de plantes vasculaires. Dans la sous-tribu des Angraecinae, le genre Angraecum, décrit comme un groupe polyphylétique (Carlsward et al., 2003; Carlsward et al., 2006; Micheneau et al., 2008a), contient près du tiers des espèces de la sous-tribu, soit 221 espèces sur les quelques 762 reconnues (Chase et al., 2015; Govaerts et al., 2015).

Angraecum Bory (1804), connu sous le nom « d’orchidée étoile » à cause de ses fleurs rappelant la forme d’une comète, est un genre restreint à l'Afrique continentale et à la région ouest de l’Océan Indien. Le centre de diversité du genre se situe à Madagascar (Hermans et al., 2007; Cribb et Hermans, 2009), où l’on dénombre 142 espèces dont 90 % sont endémiques à l’île (Govaerts et al., 2015). Ce genre comprend le fameux A. sesquipedale ou étoile de Darwin, connu pour son éperon de 30 cm de longueur et pour son papillon pollinisateur sphingidé à long proboscis, le Xanthopan morganii praedicta. Toutes les espèces appartenant à ce groupe sont ornementales et sont très convoitées en horticulture. Certaines espèces, comme A. longicalcar, sont aussi utilisées en industrie cosmétique pour leur fragrance. Mais, comme la plupart des orchidées tropicales, certaines espèces d’Angraecum se font rares en milieu naturel à cause de la perte d’habitat et diverses pressions anthropiques.

Alors que la découverte du genre Angraecum date de deux siècles, notre connaissance du groupe demeure limitée. Il faut d’abord comprendre que le grand nombre d’espèces et la diversité à l’intérieur d'Angraecum rend l’étude de ce groupe complexe (Garay, 1973; Stewart et al., 2006; Micheneau et al., 2008a). Les espèces ont été réparties en 19 sections (Garay, 1973) selon leurs ressemblances morphologiques, mais cette assignation n’est pas toujours évidente (Stewart et al., 2006). Par ailleurs, la délimitation même du genre pose problème à cause de l’homoplasie des caractères utilisés pour définir les genres dans la sous-tribu des Angraecinae. Jusqu'à maintenant les données moléculaires n’ont pas permis de résoudre entièrement la phylogénie faute d’un échantillonnage suffisant (Micheneau et al., 2008a), bien qu’elles aient semblé prometteuses. La sous-tribu des Angraecinae, à laquelle appartient le genre, paraît être un groupe polyphylétique (Carlsward et al., 2003; Carlsward et al., 2006) et le genre Angraecum lui-même pourrait aussi être artificiel (Micheneau et al., 2008a). Par

ailleurs, les données moléculaires et morphologiques ne sont pas concordantes, ce qui remet en cause la fiabilité des sections et des genres. Enfin, la plus grande diversité d’Angraecum à Madagascar suscite des questions quant aux mécanismes de diversification des espèces, à l’origine du groupe et à sa biogéographie historique.

Afin de mieux comprendre l’évolution du genre Angraecum, notre étude a trois objectifs. Le premier consiste à évaluer la diversification du genre Angraecum à Madagascar, notamment afin de vérifier si la diversité morphologique est associée à la diversité des espèces. Le deuxième consiste à évaluer la taxonomie du genre Angraecum, défini de façon monophylétique, en combinant les données moléculaires et morphologiques. Enfin, le troisième objectif consiste à évaluer les hypothèses biogéographiques, notamment celle de l'origine malgache de l'ancêtre d’Angraecum sensu stricto.

1. Historique du genre Angraecum

Etymologiquement, le mot Angraecum vient du Malais ‘angrek’, qui signifie "orchidées épiphytes" en Asie du Sud-est (Stewart et al., 2006).

Le genre Angraecum Bory (1804), dont le spécimen type est A. eburneum, fut découvert par Jean-Baptiste Geneviève Marcellin Bory de Saint Vincent lors de son voyage à la Réunion et dans les iles voisines au tout début des années 1800. Louis Marie Aubert du Petit-Thouars, connu sous le nom de Petit-Thouars, a découvert à son tour une vingtaine d’autres espèces, dont A. sesquipedale Thouars, dans les deux décennies suivantes. L’ère de l’empire colonial a favorisé la recherche et la découverte d'orchidées dans les pays tropicaux comme Madagascar. En effet, les orchidées tropicales furent un centre d’intérêt pour la société bourgeoise européenne dans les années 1800 (Arditti et al., 2012). Peu de temps après la découverte du genre Angraecum, la commercialisation des espèces avait commencé et la demande s'accrut sur le marché. Au Royaume-Uni par exemple, une plante de A. sesquipedale se vendait à 20 £ (soit 31,75 $) vers les années 1860-1870, ce qui équivaut actuellement à 1460 £ (soit 2318,13 $) (Arditti et al., 2012). L’intérêt pour Angraecum n’a jamais diminué depuis. En horticulture, on compte plus de 45 hybrides interspécifiques et 28 hybrides

intergénériques (RHS, 2015). Certaines espèces comme A. longicalcar sont maintenant utilisées dans l’industrie cosmétique à cause de leur fragrance.

Mais l'histoire la plus marquante sur le genre Angraecum reste celle de l'A. sesquipedale Thouars et de la prédiction fameuse faite par Darwin (1862) concernant son espèce pollinisatrice, et la théorie de coévolution qui en a découlé. En effet, après avoir reçu un spécimen de l’A. sesquipedale de son ami Bateman en 1862, Darwin fut émerveillé de la longueur de l’éperon de cette espèce, qui fait environ 30 cm de long. Il prédit que le papillon qui serait capable de polliniser A. sesquipedale devrait avoir une trompe aussi longue que l’éperon de la plante, alors qu’aucune espèce possédant cet attribut n'avait été découverte à l’époque. Ce n’est que 41 ans plus tard que le Xanthopan morganii praedicta, un papillon Sphingidé, fut découvert (Rothschild et Jordan, 1903), et 130 ans plus tard que preuve fut donnée de la pollinisation de l'A. sesquipedale par cet insecte (Wasserthal, 1993). Darwin avait conclu que l'A. Sesquipedale et son insecte pollinisateur (X. morganii predicta) seraient en coévolution, faisant allusion à la longueur de l’éperon de la plante qui concorde avec la longueur de la trompe de l’insecte. Cette théorie fut soutenue puis démontrée par Nilsson et al. (1985) après leurs études sur d’autres espèces d'Angraecum, A. compactum et A. arachnites, qui seraient en coévolution avec Panogena lingense. Cette théorie de coévolution reste tout de même un sujet de discussion pour les biologistes, pour la raison que ces insectes pollinisateurs visitent d’autres plantes et non seulement Angraecum (Nilsson et al, 1998; Jermy, 1999).

Sur le plan de la conservation et de la protection, le statut des espèces appartenant au genre Angraecum est mal connu faute d’investigation. Cependant, quatre espèces figurent sur la liste rouge de l’IUCN (2015): A. longicalcar (Bosser) Senghas, endémique du haut plateau de Madagascar, en danger critique; A. rubellum Bosser, endémique de la forêt humide de basse altitude de Madagascar, en danger critique; A. sanfordii P. J. Cribb et B.J. Pollard, endémique du Cameroun, en danger; et A. pyriforme Summerh., une espèce de l’Afrique de l’ouest, vulnérable. La collecte abusive pour la commercialisation est l’un des facteurs majeurs qui entraîne la rareté des espèces dans leur milieu naturel, malgré le fait que toutes les espèces d’orchidées sont classées à l'Annexe II de la CITES (2015). Mais les feux de brousse provoqués qui sévissent dans les forêts et prairies favorisent aussi la destruction, la fragmentation et la disparition de l’habitat et des espèces (Whitman et al. 2011; IUCN, 2015).

2. Classification et description d’Angraecum

Suivant l’APG III (2009), le genre Angraecum appartient à la famille des Orchidaceae Juss., sous-famille des Epidendroideae Lindl., tribu des Vandae Link, sous-tribu des Angraecinae Summerh.

Dressler (1993) avait subdivisé la tribu des Vandeae en quatre sous-tribus, Aeridinae, Aerangidinae, Angraecinae et Polystachyinae, qui regroupent 103, 36, 18 et 4 genres, respectivement. Cette tribu regroupe des espèces à croissance monopodiale (sauf les Polystachyinae qui ont une croissance polypodiale), avec des vélamens (Porembski et Barthlott, 1988; Arditii, 1992) et des graines de type Vanda (Dressler, 1993; Sin et al., 2002). Les Aeridinae et les Polystachyinae sont monophylétiques. Cependant, les Aerangidinae et les Angraecinae, qui forment le groupe informel des «angraecoides», sont tous deux polyphylétiques, mais forment ensemble un groupe naturel (Carlsward et al., 2006). Une forte ressemblance morphologique existe entre les deux groupes, leur différence se situant au niveau du gynostème (Dressler, 1993). Chez les Angraecinae, les deux pollinies sont portées par des stipes distincts attachés chacun sur un viscidium, alors que chez les Aerangidinae, les deux pollinies sont portées sur un seul viscidium (Fig. 1). Les récentes études phylogénétiques moléculaires sur la famille des Orchidaceae reconnaissent plutôt quatre sous-tribus à l’intérieur de la tribu des Vandeae: Aeridinae, Angraecinae (Aerangidinae + Angraecinae), Adrorhizinae, et Polystachyinae (Chase et al., 2015).

Le genre Angraecum, type de la sous-tribu des Angraecinae, est caractérisé par des plantes épiphytes, lithophytes, rarement terrestres; à racines aériennes généralement développées et longues; avec un port souvent dressé, pendant ou grimpant; des tiges courtes, allongées ou absentes; des feuilles distiques, généralement alternes, de forme et de longueur très variables; des entres nœud courts ou allongés; des feuilles minces, épaisses, ou charnues, de forme filiforme, lancéolée, ou ovale; inflorescences solitaires, racèmes ou épis; des fleurs petites à grandes, de couleur blanche, verte, jaune ou ocre; avec un éperon court à long; avec une anthère terminale, operculée avec des partitions réduites et deux pollinies (Dressler, 1993; Du Puy et al., 1999). Le nombre de chromosomes varie selon l'espèce et peut être 2n = 34, 36, 38, 40, 42, 44, 46, 48 ou 50 (Dressler, 1993; Micheneau, 2005).

Fig. 1. Comparaison entre Angraecinae et Aerangidinae. 1: colonne avec anthère; 2: rostellum; 3: pollinies (Dressler, 1993).

3. Biologie, écologie et répartition d’Angraecum

Le genre Angraecum s’étend de l’Afrique occidentale à l'Afrique orientale (subsaharienne), et dans l’Océan Indien à Madagascar, dans les Mascareignes (Réunion, Maurice, Rodrigues), dans les iles Comores, aux Seychelles et au Sri Lanka. Plus de 95% des espèces appartenant au genre sont épiphytes (Stewart et al., 2006; Hermans et al., 2007). Grace à leur système racinaire aérien et à leurs feuilles souvent charnues ou épaisses, les espèces présentent une plasticité et une tolérance au stress hydrique (Kerbauy et al., 2012), mais elles demeurent sensibles au feu (Whitman et al., 2011). D’après les études menées par Kluge et al. (1997), les espèces appartenant au genre Angraecum ont deux types de mécanismes photosynthétiques: CAM et C3. Cette étude a montré que les espèces épiphytes sont de type CAM, ce qui est confirmé par les travaux de Kerbauy et al. (2012). Comme toutes les orchidées, Angraecum développe une relation symbiotique avec des champignons

mycorhiziens (Arditti, 1992). Aucune étude sur la spécificité de ces microorganismes avec notre groupe n’a encore été faite, mais l’on sait qu’ils assurent un apport en carbone à la plante (Otero et Flanagan, 2006). Les espèces d'Angraecum croissent généralement à une altitude de 0 à 2000 mètres dans les régions tropicales humides, subhumides ou sèches (Du Puy et al., 1999; Stewart et al., 2006; Hermans et al., 2007). Elles abondent en forêt tropicale humide où la dénivellation est assez élevée. Des températures entre 10°C et 16°C, et une précipitation supérieure à 2000 mm par an sont considérées comme des conditions favorables au développement des orchidées épiphytes (Acharya et al., 2011).

Angraecum est considéré comme un genre sphingophile, c'est-à-dire présentant un mécanisme d’adaptation à la pollinisation par les papillons de nuit, de par la couleur claire des fleurs, des éperons nectarifères souvent longs et effilés, et l’émission de fragrance à la tombée de la nuit (Johnson et Steiner, 2000). Bien que des preuves de pollinisation par les papillons de nuit existent pour A. sesquipedale (Wasserthal, 1993, 1997), A. arachnites et A. compactum (Nilsson et al., 1985), la connaissance des espèces pollinisatrices d’Angraecum reste encore réduite. De récentes études avaient montré que des agents pollinisateurs autres que les papillons existent (Micheneau, 2005; Micheneau et al., 2008c, 2009, 2010). Ainsi, des observations sur la visite de l'A. bracteosum par l’oiseau Zosterops borbonicus et de l'Angraecum cadetii par le criquet Glomeremus orchidophilus ont été faites. Suite à ces observations, Micheneau et al. (2009) avaient conclu que l’adaptation de la plante suite à un changement d’habitat ou à la colonisation d’un nouvel habitat pourrait conduire à l’apparition de nouveaux agents pollinisateurs.

4. Choix du site d’étude et échantillonnage

Madagascar a été choisi comme site d’étude à cause de sa grande richesse en biodiversité, mais surtout parce que c’est le centre de diversité de notre groupe (Stewart et al., 2006). La diversité floristique de l’ile s'explique par les différentes formations géologiques qui constituent le territoire malgache (Du Puy et Moat, 1996). Mais chez les plantes épiphytes comme les orchidées, la diversité est aussi conditionnée par d’autres facteurs comme le gradient d’élévation et le climat (Acharya et al., 2011).

Fig. 2. Carte de végétation de Madagascar (Du Puy et Moat, 1996)

Humbert (1955) avait subdivisé Madagascar en sept zones de végétations: domaine de l’est (0 – 800 m), domaine du Sambirano (0 – 800 m), domaine du centre pentes orientales (800 – 2000 m), domaine du centre pentes occidentales (800 – 2000 m), domaine des hautes montagnes (2000 m et plus), domaine de l’ouest (0 – 800 m) et domaine du sud (800 – 2000 m). Plus tard, Du Puy et Moat (1996) ont raffiné les zones de végétation (Fig. 2) en se basant sur la description d’Humbert (1955) et celle de Faramalala (1995). La végétation primaire à Madagascar peut se résumer comme suit: à l’est, une forêt dense humide (littorale, basse et moyenne altitude, montagne), et à l’ouest, une forêt dense sèche caducifoliée, une savane et une forêt de palétuvier. Le sud est dominé par une forêt dense sèche et épineuse, alors que le centre est marqué par une forêt de montagne avec buisson éricoïde et une formation d’inselbergs. Le climat de l’île se subdivise comme suit: au centre, un climat tropical de montagne, généralement froid avec précipitation annuelle inférieure à 1500 mm; à l’est, un climat tropical humide avec précipitation annuelle supérieure à 2000 mm conditionné par le régime d’alizé sur toute la partie est et de mousson dans la partie nord de l’île; au sud, un climat semi-aride soumis au vent du sud (chaud et sec); enfin à l’ouest un climat tropical sec (Donque, 1975).

Une série d’échantillonnages a été réalisée de 2007 à 2012 dans 12 sites différents de Madagascar incluant toutes les zones de végétations existantes à l’exception du centre pente occidentale et du sud car elles comptent peu d’espèces dont, de surcroît, les données de séquences sont déjà publiées (Carlsward et al., 2006; Micheneau et al., 2008). Les caractéristiques de chaque site sont présentées au Tableau 1. Les expéditions ont été faites en collaboration avec le Madagascar Biodiversity Partnership (une organisation non- gouvernementale) et l’Université d’Antananarivo. Une partie des feuilles (2 cm x 2 cm) a été collectée puis conservée dans du gel de silice (Chase et Hills, 1991) pour les études moléculaires. Chaque individu échantillonné a été photographié et ses coordonnées géographiques enregistrées. Des spécimens d’herbiers ont été collectés pour les sites qui n’avaient pas le statut d’aire protégée, en accord avec la politique de conservation et la gestion locale des aires protégées. Au total, nous avons récolté 727 échantillons issus d’une centaine d’espèces et 183 spécimens d’herbier. Les spécimens d’herbiers seront déposés à l’herbier de Tsimbazaza à Antananarivo (TAN) et à l’herbier Marie-Victorin (MT). Afin de compléter

notre échantillonnage, nous avons utilisé les collections du Jardin Botanique de Montréal (don de M. Lecoufle).

Tableau 1. Liste et caractéristiques des sites d’étude.

Sites d’étude date de région type de latitude altitude l’expedition formation longitude (m) Ambohitantely 23 au 27 Avril Analamanga Forêt humide de -18,19580556 1529 à 2012 montagne 47,28955556 1639 Angavokely 18 au 22 Avril Analamanga Forêt humide de -18,92172222 1569 à 2012 montagne 47,73122222 1692 Anjozorobe 3 au 8 Mai Analamanga Forêt humide de -18,39330556 1550 à 2012 montagne 47,93930556 1612 Bemaraha 1 au 5 Déc Melaky Forêt sèche -19,04461111 45 à 2007 caducifoliée 44,78047222 162 Lakia 28 au 31 Janv Vatovavy- Forêt humide de -21,49475000 193 à 2008 Fitovinany basse altitude 47,89972222 444 Mananara 14 au 22 Fév Analanjirofo Forêt humide de -16,31261111 5 à Nord 2008 basse altitude 49,78536111 433 Manongarivo 10 au 19 Déc Diana Forêt humide de -14,02302778 415 à 2008 moyenne altitude 48,25522222 835 Mantadia 10 au 15 Janv Alaotra- Forêt humide de -18,82641667 980 à 2008 Mangoro moyenne altitude 48,44086111 1149 Marojejy 10 au 23 Mars Sava Forêt humide de -32,98469444 779 à 2009 moyenne altitude 92,00141667 865 Sangasanga 4 au 14 Sept Vatovavy- Forêt humide de -21,37461111 106 à 2009 Fitovinany basse altitude 47,86588889 404 Torotorofotsy 15 au 19 Déc Alaotra- Forêt humide de -18,78208333 925 à 2007 Mangoro moyenne altitude 48,43369444 1029 Vatovavy 15 au 30 Sept Vatovavy- Forêt humide de -21,40777778 338 à 2009 Fitovinany basse altitude 47,94100000 591

5. Systématique et phylogénie d’Angraecum

Les 19 sections de Garay (1973), basées sur des caractères morphologiques, servent présentement de référence pour l’étude du genre Angraecum (Tableau 2). En fait, le travail de Garay résulte de la révision systématique des travaux effectués par Schlechter (1918, 1925), Perrier de la Bathie (1941), et Summerhayes (1958). Stewart et al. (2006) ont essayé de déplacer certaines espèces d’une section à une autre en conservant les sections de Garay et en

Tableau 2. Liste des 19 sections de Garay (1973) avec leur répartition géographique; les chiffres donnent le nombre d'espèces de chaque région et ceux entre paranthèses les espèces endémiques (Govaerts et al., 2015). Les sous-espèces et variétés ne sont pas incluses. Abréviations: Af: Afrique; E: est; C: centrale; O: ouest; Com: Comores; Mad: Madagascar; Msc: Mascareignes; Sey: Seychelles; Sri: Sri Lanka.

Sections Af-O Af-C Af-E Com Mad Msc Sey Sri Total Acaulia 7(7) 7 Afrangraecum 18(10) 9(1) 9(5) 24 Angraecoides 1(1) 7(6) 3(2) 10 Angraecum 1 2 16(15) 2(1) 1 17 Arachnangraecum 1(1) 2(1) 9(7) 4(3) 14 Boryangraecum 1(1) 4(3) 1 10(9) 2(1) 16 Chlorangraecum 2(2) 2 Conchoglossum 1(1) 4(4) 3(3) 8 Dolabrifolia 3(3) 3 Filangis 7(7) 1(1) 8 Gomphocentrum 1(1) 1 15(13) 5(3) 3(1) 1 19 Hadrangis 4(4) 4 Humblotiangraecum 1 5(4) 5 Lemurangis 1(1) 5(4) 2(1) 7 Lepervenchea 6(6) 1(1) 7 Nana 4(4) 10(9) 6(5) 19 Pectinaria 5(3) 2 5(4) 1 10 Perrierangraecum 2(2) 30(30) 2(2) 34 Pseudojumellea 4(2) 3(1) 5 Inconnu 1(1) 1(1) 2 Total 29(19) 11(1) 27(21) 7(1) 142(129) 37(27) 4(1) 1 221

créant des sous-groupes au sein des sections, ce qui rejoint un peu la subdivision des sections de Perrier de la Bathie (1941). Toutefois, les différences morphologiques sont tellement grandes au sein des sections mêmes qu’il leur était difficile de faire consensus sur la section où ils allaient placer certaines espèces. D'un autre côté, les résultats phylogénétiques basés sur des données moléculaires ne reflètent pas les sections et il en ressort que le genre Angraecum est polyphylétique (Carlsward et al., 2003; Micheneau et al., 2008a; Simo-Droissart et al., 2013).

6. Diversification chez Angraecum

La particularité du genre Angraecum est sa grande diversité morphologique (Fig. 3). Les sections ont été établies en premier lieu sur la base des différences morphologiques et florales entre les espèces (e.g., Schlechter, 1925; Perrier de la Bathie 1941; Garay, 1973). Par exemple, la section Nana réunit toutes les petites plantes ayant de petites fleurs, comme Angraecum nanum, qui mesure à peine 3 cm de haut et produit des fleurs de quelques milimetres de diamètre. La section Angraecum, quant à elle, regroupe les grandes plantes pouvant atteindre jusqu’à 1 m de hauteur avec de grandes fleurs, comme chez Angraecum longicalcar. La taille des plantes est souvent corrélée avec celle des fleurs et la longueur de l’éperon, mais ceci n’est pas le cas pour toutes les espèces, ce qui pourrait créer une certaine ambiguité entre l’ancien système de classification basé sur la morphologie et celui basé sur les données moléculaires. Certes, la diversité morphologique chez Angraecum constitue un outil important pour évaluer la systématique du groupe, mais elle peut aussi être efficace pour évaluer leur histoire évolutive. Plusieurs études utilisent maintenant la diversification d’un groupe de taxons pour évaluer les mécanismes de spéciation (e.g. Janssen et al., 2008; Anthony et al., 2010; Jonsson et al., 2012; Christidis et al., 2014). Par exemple, Freudenstein et Chase (2015) ont démontré que le passage vers l’épiphytisme chez les orchidées est un caractère innovateur ayant conduit à la grande diversification de la sous famille des Epidendroidae.

Fig. 3. Diversification du genre Angraecum. (A) A. sedifolium, (B) A. danguyanum, (C) A. mirabile, (D) A. longicalcar, (E) A. sesquipedale (Photo: Andriananjamanantsoa H.).

7. Biogéographie historique d’Angraecum

La biogéographie « étudie les organismes vivants, les plantes et les animaux, à la surface du globe, dans leur répartition, dans leur groupement et dans leurs relations avec les autres éléments du monde physique et humain » (Elhaï, 1968; Galochet, 2008). D’une manière simplifiée, la biogéographie est l’étude de la répartition géographique des organismes (Crisci, 2001). Le botaniste suisse de Candolle (1820) fut le premier à distinguer entre biogéographie écologique et historique. Selon lui, la biogéographie écologique s'intéresse aux processus écologiques qui se produisent sur une courte période et selon une petite échelle spatiale, tandis que la biogéographie historique s'intéresse aux processus évolutifs au cours des millions d'années à une grande échelle, souvent mondiale.

En biogéographie historique, trois types de processus qui modifient la disposition géographique et spatiale des organismes ont été identifiés: l'extinction, la dispersion et la vicariance. Ces processus sont maintenant évalués suivant trois éléments majeurs considérés comme influents dans l’étude de la biogéographie historique: la tectonique, la cladistique et la perception des biologistes de la biogéographie (Crisci, 2001). Pour l’étude de la biogéographie historique de notre groupe, nous allons essayer de trouver l’origine et les causes de la répartition des taxons en utilisant les données géologiques, écologiques, biologiques, et phylogénétiques.

D’après les données géologiques, Madagascar se serait séparé de l’Afrique à partir du bassin somalien du côté nord-ouest, puis du bassin du Mozambique du côté sud-ouest. Cette fracturation aurait eu lieu il y a environ 120 Ma (Eagles et Konig, 2008). La séparation entre Madagascar et l’Inde se serait produite il y a environ 80 Ma (Fig. 4), et la séparation entre l’Antarctique et l’Australie entre 40-55 Ma (Aslanian et Moulin, 2010; Torsvik et al., 2010). Cette séparation est aussi corroborée par des données biologiques, notamment fossiles (Ali et Krause, 2011). L’hypothèse de ponts reliant le bloc Indo-Madagascar avec le reste du Gondwana pendant le mi-Aptien a été avancée (Ali et Aitchison, 2009), mais cela reste controversé (Ali et Krause, 2011). L’ordre chronologique de la fragmentation du bloc de Gondwana-Est (Madagascar, Inde, Australie, Antarctique) reste un débat ouvert (Aslanian et Moulin, 2010; Torsvik et al., 2010).

Deux hypothèses majeures sont avancées sur l’origine des espèces à Madagascar, la dispersion transocéanique (Schatz, 1996; Yoder et Nowak, 2006) et la vicariance gondwanienne (Leroy, 1978; Grubb, 2003; Yoder et Nowak, 2006). L’origine vicariante des espèces malgaches est discutable à cause de l’âge géologique du pays (Bremer et al. 2004; Janssen et Bremer, 2004; Anderson et al., 2005; Davis et al., 2005). La colonisation par des espèces africaines pendant le Cénozoïque est l’hypothèse la plus soutenue (Yoder et Nowak, 2006). Une telle hypothèse est appuyée par Strijk et al. (2012) dans leur étude sur la dispersion de Psiadia (Asteraceae). Par ailleurs, de recentes études soutiennent une hypothèse sur l’origine Asiatique de plusieurs lignées malgaches (Cheke et Hume, 2008; Warren et al., 2010; Krüger et al., 2012; Buerki et al., 2013), remettant en question l’origine africaine des lignées

malgaches et suggérant d’autes possibilités de dispersion à longue distance en direction est- ouest.

Fig. 4. Position approximative de Madagascar dans le bloc du Gondwana-Est au Crétacé. Marges continentales interprétées à partir des données gravimétriques satellitaires (Eagles et Konig, 2008).

Pour la biogéographie historique d’Angraecum, seuls Micheneau et al. (2008a) ont émis une hypothèse sur l’origine biogéographique des espèces dans leurs travaux sur la phylogénie et la biogéographie des angraecoides des Mascareignes. Cette étude, basée sur l’analyse des aires ancestrales de Bremer (1992) proposait une origine malgache des Angraecinae (Micheneau et al., 2008a) qui serait apparue il y a environ 21 Ma (Micheneau et al., 2010). Plusieurs auteurs ont avancé différents âges pour la famille des Orchidaceae, pour certains elle serait jeune à 26 Ma (Wikstrom et al., 2001) ou 40 Ma (Bremer, 2000), alors que

pour d’autres elle aurait apparue il y a environ 110 Ma (Janssen et Bremer, 2004). Ce n’est que récemment que deux fossiles appartenant à la sous-famille des Orchidoideae et à la sous- famille des Epidendroideae ont été découverts en Nouvelle-Zélande (Conran et al., 2009), ayant permis d’estimer l’âge de la famille à environ 76–84 Ma (Ramirez et al., 2007). Ceci a été par la suite validé par Gustafsson et al. (2010) après avoir integré d’autres fossiles d’orchidées dans leurs analyses.

8. Problématiques, objectifs et hypothèses de recherche

Compte tenu des études antérieures sur la systématique d’Angraecum (Carlswarde et al., 2006; Micheneau et al., 2008; Simo-Droissart et al., 2013), plusieurs questions demeurent concernant la monophylie du genre et des sections sensu Garay (1973), et la position taxonomique des espèces (délimitation du genre). La résolution partielle de la phylogénie, potentiellement due à un faible échantillonnage (Micheneau et al., 2008), limite d’autres études, comme celle de l’évolution et de la diversification des espèces ou encore celle de l’histoire biogéographique qui nécessitent une phylogénie entièrement résolue. Ces études pourraient fournir des éléments essentiels qui permettent d’évaluer le statut de conservation des espèces dans le but de mieux établir une politique de conservation basée sur le principe de nécessité et de priorité.

Face à ces problématiques, mon projet de doctorat comporte quatre objectifs majeurs relatifs à la systématique évolutive et la biogéographie. Le premier objectif (chapitre 1) consiste à reconstruire la phylogénie d’Angraecum en augmentant le nombre d'espèces provenant de Madagascar qui manquaient dans les études antérieures et en ajoutant des caractères morphologiques aux analyses. Avec les 727 échantillons que nous avons collectés, nous espérons accroître la résolution de la phylogénie. Le deuxième objectif (chapitre 2) consiste à faire une révision systématique du genre en se basant sur les résultats obtenus dans le premier chapitre. Le troisième objectif (chapitre 1 et 3), consiste à évaluer la diversification d’Angraecum et de la sous-tribu des Angraecinae, notament l’effet des traits floraux (chapitre 1) et la stratification forestière (chapitre 3) sur la diversification des espèces, en utilisant la phylogénie moléculaire et les caractères morphologiques. Le quatrième objectif (chapitre 3)

consiste à évaluer l’histoire biogéographie d’Angraecum et des Angraecinae afin de comprendre l’origine, le sens de la migration ou encore les mécanismes de dispersion des taxons.

Afin de répondre à nos objectifs, cinq hypothèses sont avancées:

Hypothèse 1: un Angraecum monophylétique devrait exclure les sections africaines;

Hypothèse 2: la phylogénie infirme la classification sectionnelle de Garay (1973) pour Angraecum;

Hypothèse 3: certains caractères morphologiques utilisés en taxonomie, notamment floraux, furent impliqués dans la diversification d’Angraecum;

Hypothèse 4: la diversification d’Angraecum s’est faite par radiation adaptative rapide;

Hypothèse 5: l’ancêtre d’Angraecum s. s. est malgache, et la dispersion s’est effectuée de Madagascar vers d’autres régions.

Chapitre 1: Diversification of Angraecum (Orchidaceae, Vandeae) in Madagascar: species accumulation through time rather than rapid radiation

Andriananjamanantsoa H. Notahianjanahary1, S. Engberg2, E. Louis2, 3 and L. Brouillet1

1. Département de sciences biologiques, Université de Montréal, 90 Av. Vincent d’Indy, Montréal (QC) H2V 2S9, Canada

2. Omaha’s Henry Doorly Zoo and Aquarium, 3701 S 10th St, Omaha (NE) 68107, USA

3. Madagascar Biodiversity Partnership, VO18bis Manakambahiny, Antananarivo-101, Madagascar

En préparation. Journal ciblé: Taxon (Modifié pour les besoins de la présente thèse)

Tous les auteurs ont donné leur autorisation pour ce manuscrit à être inclus dans cette thèse

Contribution respective des auteurs:

Tahiana Andriananjamanantsoa: Conception du projet, collecte des données sur terrain, traitement et analyse des données au laboratoire, interprétation et rédaction

Shannon Engberg: Supervision en laboratoire, conseils et recommandations

Edward Louis Jr.: Supervision sur terrain, conseils, recommandations et co-direction

Luc Brouillet: Supervision, conception du projet, conseils, recommandations, corrections et direction générale

1.1. Résumé/ Abstract

Résumé

Madagascar fait partie des 34 points chauds de biodiversité dans le monde. Les îles comme Madagascar sont souvent utilisées pour étudier les processus de diversification en raison de leur forte endémicité et de leur isolement. Angraecum est le plus grand genre de la sous-tribu Angraecinae (Orchidaceae) avec environ 221 espèces. Madagascar est le centre de diversité du genre avec ca. 142 espèces, dont 90% sont endémiques. La grande diversité morphologique associée à la diversification des espèces du genre à Madagascar offre des informations précieuses pour les études macroévolutives. Des phylogénies sur Angraecinae ont été publiées, mais le manque d'échantillonnage limite leur résolution, laissant des incertitudes taxonomiques. Nous présentons une nouvelle phylogénie d'Angraecum fondée sur des données de séquences chloroplastiques (matK, rps16, trnL), nucléaire (ITS2) et 39 caractères morphologiques de 194 individus des Angraecinae dont 69 nouvellement échantillonnés, incluant 98 espèces du genre Angraecum et cinq extra groupes des sous-tribus Aeridinae et Polystachyinae. En utilisant cette phylogénie, nous avons évalué les sections d'Angraecum tel que défini par Garay et étudié les modèles de diversification des espèces au sein du genre. Nous avons utilisé la méthode de parcimonie et les analyses bayésiennes pour générer des arbres phylogénétiques et dater la phylogénie. Nous avons analysé les modèles de diversification au sein des Angraecinae et d’Angraecum, en nous basant sur quatre traits floraux (couleur et taille des fleurs, position du labellum, longueur de l’éperon), en utilisant des modèles macroévolutifs afin de trouver quels traits ou états de caractères seraient associés au taux de spéciation, et reconstruit les états ancestraux des caractères. L'analyse phylogénétique a montré la polyphylie du genre Angraecum s.l. et de toutes ses sections, à l’exception de la section Hadrangis, et la compatibilité de la morphologie avec la phylogénie. La position supérieure et inférieure du labelle est le caractère principal qui permet de délimiter des clades monophylétiques au sein d'Angraecum s.s. Ce caractère semble également influer sur le taux de spéciation chez Angraecum. Le model macroévolutif basé sur la phylogénie a échoué à détecter des changements de diversification pouvant être associés à la diversification morphologique. La diversification d'Angraecum résulte d’une accumulation progressive

d'espèces à travers le temps plutôt qu’une radiation rapide, un modèle de diversification souvent rencontré dans les forêts tropicales humides.

Mots-clés: accumulation des lignées, Angraecum, taux de diversification, macroévolution, Madagascar, phylogénie

Abstract

Madagascar is one of the 34 World biodiversity hotspots. Islands like Madagascar are often used to study diversification processes because of their high endemism and isolation. Angraecum is the largest genus of subtribe Angraecinae (Orchidaceae) with about 221 species. Madagascar is the center of the diversity for the genus with ca. 142 species, of which 90% are endemic. The great morphological diversity associated with species diversification in the genus on the island of Madagascar offers valuable insights for macroevolutionary studies. Phylogenies of the Angraecinae have been published but lack of sampling and the limited taxonomic resolution limit their uses for macroevolutionary studies. We present a new phylogeny of Angraecum based on chloroplast sequence data (matk, rps16, trnL), nuclear ribosomal (ITS2) and 39 morphological characters from 194 Angraecinae specimens of which 69 were newly collected, including 98 Angraecum species and five outgroups from subtribes Aeridinae and Polystachyinae. Using this phylogeny, we evaluated the monophyly of the sections of Angraecum as defined by Garay and investigated the patterns of species diversification within the genus. We used maximum parsimony and bayesian analyses to generate phylogenetic trees and dated divergence times of the phylogeny. We analyzed diversification patterns within Angraecinae and Angraecum with an emphasis on four floral traits (flower color, flower size, labellum position, spur length) using macroevolutionary models to evaluate which traits or character states are associated to speciation rates, and inferred ancestral states of these characters. The phylogenetic analysis showed the polyphyly of Angraecum sensu lato and of all Angraecum sections except sect. Hadrangis, and that morphology is consistent with the phylogeny. The uppermost and lowermost position of the labellum was the main character helping to delimit clades within a monophyletic Angraecum sensu stricto. This character appeared also to be associated with speciation rates in

Angraecum. The macroevolutionary model-based phylogeny failed to detect shifts in diversification that could be associated directly with morphological diversification. Diversification in Angraecum resulted from gradual species accumulation through time rather than from rapid radiation, a diversification pattern often encountered in tropical rain forests.

Keywords: accumulation of lineages through time, Angraecum, diversification rate, macroevolutionary, Madagascar, phylogeny

1.2. Introduction

Madagascar is known for its biodiversity richness (Goodman and Benstead, 2005), and is a focus for evolutionary biologists who study the cause of species diversification. Many studies have been conducted using the high endemicity and diversity found in this isolated island as a model of diversification process in various taxa (Martin, 1972; Janssen et al., 2008; Townsend et al., 2009; Vences et al., 2009; Anthony et al., 2010; Jonsson et al., 2012; Reddy et al., 2012; Rakotoarinivo et al., 2013; Christidis et al., 2014). Most of these studies came to the conclusion that species diversification resulted from rapid radiation. High morphological variation associated with low genetic divergence appears to be a signature of the diversification processes operating on the island of Madagascar (Reddy et al., 2012) as well as on many other islands such as New Zealand, Hawaii, or Caribbean (Goldberg et al., 2014, Goodman et al., 2015, Lewis et al., 2015). With the improvement of methods for macroevolutionary studies (Pagel, 1994; Pybus and Harvey 2000; Maddison et al., 2007, Fitzjohn et al., 2009; Rabosky et al., 2014) it has been advocated that species diversification is not necessarily linked with adaptive radiations (Schluter, 2000), but could be a result of a gradual accumulation of ancestral lineages through time (Couvreur et al., 2011). The problems associated with studies on diversification are primarily due to insufficient data such as incomplete taxon samplings or phylogenetic uncertainties (Fitzjohn et al., 2009; Rabosky et al., 2013). Macroevolutionary studies require a good knowledge of paleontological events and eventually a fossil record to help calibrate the phylogenies, but most of the time these are lacking, making the interpretation of evolutionary histories controversial. Understanding the diversification processes of a group of organisms could be useful for biodiversity conservation

(Scantlebury et al., 2013), especially for hotspots like Madagascar (Myers et al., 2000) where priority should be given to the most valuable species due to lack of conservation funds.

Genus Angraecum is the second largest group of orchids in Madagascar; it has a high morphological diversity, which makes it a good candidate for macroevolutionary studies. Angraecum includes approximately 221 species (Chase et al., 2015; Govaerts et al., 2015) distributed from sub-Saharan Africa to Madagascar, the Indian Ocean Islands (IOI: Comoros, Mauritius, Reunion, Rodrigues, and Seychelles), and Sri Lanka. Madagascar is the center of diversity for the genus with ca. 142 species of which 90% are endemic. The majority of species are epiphytic, but some are lithophytic. Epiphytic plants are found in tropical rain forest across the eastern slope of the mountains or in mesic forests in the central highland, while lithophytic species are found on inselbergs or on limestone. Recent molecular phylogenetic work (Chase et al., 2015) placed Angraecum in subtribe Angraecinae, the latter sister to subtribe Aeridinae. The phylogenetic reconstructions of the tribe Vandeae (Carlsward et al., 2006) revealed the polyphyly of Angraecinae sensu Dressler (1993) and of many of its genera including Angraecum (Carlsward et al., 2006; Micheneau et al., 2008a). Accordingly, a new circumscription has been proposed by merging Angraecinae and Aerangidinae into a single subtribe Angraecinae (Micheneau et al., 2008a), which includes 47 genera and ca. 762 species (Chase et al., 2015). Even though revisions were made to resolve the polyphyly of several genera (e.g. Cribb et al., 2007; Cribb and Carlsward, 2012), similar attempts with Angraecum proved difficult because of limited sampling (Carlsward et al., 2006; Micheneau et al., 2008a). The study of Angraecum is complicated because of the large number of species associated with a great morphological diversity, and because of morphological similarities between many members of subtribe Angraecinae. Garay (1973) proposed 19 sections to accommodate the species according to morphological descriptions, essentially floral (e.g. flower color, flower size, spur length), as was done by previous authors (e.g., Schlechter, 1918, 1925; Perrier de la Bathie, 1941; Summerhayes, 1958). Recent molecular phylogenic work revealed that Garay’s sections are polyphyletic (Micheneau et al., 2008a; Simo-Droissart et al., 2013). Some authors proposed to remove all species that cause the polyphyly with the aim of making Angraecum monophyletic (Szlachetko and Romowicz, 2007; Simo-Droissart et al., 2013; Szlachetko et al., 2013). All strictly African sections were removed from Angraecum

sensu lato (Szlachetko et al., 2013) and elevated to generic rank (Angraecoides, Dolabrifolia, Eichlerangraecum and Pectinariella). Micheneau et al. (2008a) showed that the morphology- based classification did not reflect the molecular phylogeny, and concluded that most of the sections of Garay (1973) were non-monophyletic. This study was based on Mascarene species and was lacking samples from Madagascar.

Although, the evolution of pollinia, adaptation to epiphytic habitat, and development of crassulacean acid metabolism (CAM) photosynthesis was shown to have been involved in the diversification of the Orchidaceae (Givnish et al., 2015), little is known about the relationship between morphological variation and species diversification. With the morphological diversity observed in Angraecum, we were interested in whether the traits used to define the sections, essentially flower color, flower size and spur length, were also involved in species diversification. Questions arose also on how the diversification of the genus occured. The main objectives of the current study are to (1) reconstruct the phylogeny of Angraecum sensu Garay (1973) using a larger sample of Malagasy species and most available Angraecinae species; (2) test the monophyly of morphologically defined sections; and (3) evaluate whether species diversification corroborates with morphological diversification. Here, we present a comprehensive phylogenetic reconstruction of Angraecum sensu Garay based on molecular DNA sequence data and a more comprehensive sample of species. Using this phylogenetic framework combined with morphological data, we attempt to resolve the systematic problems existing at the sectional level, and try to understand the patterns of species diversification in Angraecum. To avoid confusion, we will use Angraecum to designate the sensu stricto genus, and Angraecum sensu lato to identify the widest concept of the genus as defined by Garay.

1.3. Material and methods

1.3.1. Taxon sampling

Plant tissue for DNA extraction was obtained from field collected silica gel dried samples (Chase and Hills, 1991) of most of Malagasy Angraecum s.l. and Angraecinae species. The remaining materials, in the form of sequences data, came from previous studies

(Carlsward et al., 2006; Micheneau et al., 2008a; Rakotoarivelo et al., 2013; Simo-Droissart et al., 2013). A total of 194 specimens of which 69 were newly sampled were included in our analyses, comprising 98 Angraecum s.l., 17 Jumellea, 10 Aeranthes, and 40 other group genera (Appendix 1). Acampe ochracea, Aerides odorata, Phalaenopsis cornu-cervi, Vanda tricolor (subtribe Aeridinae), and Polystachya fulvilabia (subtribe Polystachyinae) were used as outgroups based on previous phylogenetic studies (Carlsward et al., 2006; Micheneau et al., 2008a; Chase et al., 2015). Because of conservation policy in Madagascar National Parks, the collection of herbarium specimens was not allowed in protected areas. Therefore, pictures were taken to serve as vouchers and were stored at the Marie-Victorin Herbarium (MT), which eventually will be accessible via Canadensys (in progress). Elsewhere, voucher specimens were collected and will be deposited at the national herbarium of Madagascar (TAN) and MT. Overall, 32 of the 69 sampled specimens were vouchered (Appendix 1).

1.3.2. Morphological data

To investigate morphological evolution and diversification pattern, 39 characters (13 vegetative, and 26 floral) were scored for the taxa represented in our molecular sampling. Type specimens preserved in the herbaria at Kew (K) and Paris (P), where over 80% of holotypes are located, were scored for each Angraecum s.l. species. Missing characters were documented from literature descriptions (Perrier de la Bathie, 1941; Stewart et al., 2006; Cribb and Hermans, 2009). ). For type specimens preserved in other herbaria (e.g. B, BM, MO) that we did not borrow and for other genera, characters were scored from photographs of living material and voucher specimens as well as from literature descriptions (Stewart et al., 2006; Cribb and Hermans, 2009). Character descriptions and other background information are provided in Appendix 2, while the generated matrix is presented in Appendix 3.

1.3.3. PCR amplification and DNA sequencing

Total DNA was extracted from 20–30 mg of silica-dried gel leaf material following the modified hexadecylmethylammonium bromide (2x CTAB, 2% (w/v)) extraction protocol of

Doyle and Doyle (1987); 1% polyvinylpyrrolidone (PVP) and 0.2% of β-mercaptoethanol was added to the total volume of the extraction buffer. Three plastid DNA markers were amplified: matK coding gene, rps16 intergeneric spacer, and trnL intron. The amplification of the matK region was performed using the barcoding primers 472F/1248R designed by Yu et al. (2011). The rps16 region was amplified using the primers 1F/2R designed by Oxelman et al. (1997). The trnL intron was amplified using the primers 49873F/50272R designed by Taberlet et al. (1991). The nuclear ribosomal internal transcribed spacer ITS2 region was amplified using the primers S2F/S3R designed by Chen et al. (2010). The PCR reactions contained, 1X PCR reaction buffer, 2.5 mM MgCl2, 0.16 µM of each primer, 0.2 mM of each dNTP, 0.4 % bovine serum albumin (BSA), 2 units of Taq DNA polymerase, 30 ng of template DNA, adjusted to a final volume of 25 µL with de-ionized water. PCR conditions are the same as described in Yu et al. (2011). PCR ampliﬁcations were performed on a GeneAmp PCR System 9700 thermocycler (Applied Biosystems, Foster City, CA), and resulting PCR products were puriﬁed using exonuclease I and shrimp alkaline phosphatase (ExoSAP; Silva et al., 2001). The purified products were cycle sequenced using a BigDye® terminator sequencing kit (Life Technologies, Carlsbad, CA). Sequences were analyzed with an Applied Biosystems 3130xl genetic analyzer at the Omaha’s Henry Doorly Zoo and Aquarium (NE, USA). Sequence fragments were aligned to generate a consensus sequence using Sequencher® 4.10 (Gene Codes Corporation; Ann Arbor, MI). Sequence data were also obtained from the Canadian Center for DNA Barcoding (part of matK and ITS2) and from GenBank for previous studies (Carlsward et al., 2006; Micheneau et al., 2008a; Rakotoarivelo et al., 2013; Simo-Droissart et al., 2013). Following automatic alignment using SeaView v4 (Gouy et al., 2010), alignments were edited manually using BioEdit v7.1.3 (Hall, 1999). All newly generated sequences have been deposited in GenBank (Appendix 1), and the matrices in TreeBase (in progress).

1.3.4. Phylogenetic analyses

Four matrices were produced: (1) combined plastid, (2) nuclear ribosomal, (3) morphological matrix, and (4) combined plastid and morphological matrix. They were analyzed using maximum parsimony (MP) and Bayesian analyses (BA). Tree searches under

parsimony were conducted in PAUP* version 4.0 Beta (Swofford, 2002). A preliminary heuristic search was performed with 1000 replicates of random addition sequence, tree bisection–reconnection (TBR) branch swapping, retaining twenty most parsimonious trees at each replicate. Starting with the trees kept in memory from this initial analysis, a second heuristic search was performed with TBR and 10000 trees were saved. A strict consensus tree was constructed for each analysis. Branch support was estimated using 5000 bootstraps replicates under a heuristic strategy with one random addition-sequence replicate TBR branch swapping. Bayesian analyses were performed with MrBayes v3.1.2 (Huelsenbeck and Ronquist, 2003). The best nucleotide substitution model was selected with jModelTest2 (Posada and Crandall, 1998) using the Akaike information criteria (AIC) (Akaike, 1974). For all regions, the GTR+I+G model scored best and was selected. For the combined molecular and morphological matrix, data were partitioned as DNA and standard respectively. Two parallel runs of eight Metropolis Coupled Markov Chain Monte Carlo (MCMCMC) each, and four swaps per swapping cycles for 15 million generations were undertaken. Trees were sampled every 1000 generations, and the first 25% generations were discarded as burnin. The 50% majority consensus tree with Bayesian clade credibility was built from post-burnin trees.

1.3.5. Estimation of divergence times

Divergence times were estimated using a relaxed molecular clock approach as implemented in BEAST (Drummond and Rambaut, 2007). The combined plastid matrix was used as input data in BEAUti. The GTR+G+I model was selected as substitution model. A relaxed lognormal molecular clock model was selected. The Yule model was selected as tree prior. The age for the root of the tree was set to a normal distribution with mean 35 Ma and a standard deviation of 3 (giving a 95% CI ranging from 30.07 – 39.93 Ma). Because fossil data are rare in the Orchidaceae (Iles et al., 2015.) only three having been recorded so far, none of them close to our group we used the age estimate of Phalaenopsis (35 Ma) found by Gustafsson et al. (2010) to calibrate the stem root of Vandeae. The prior distribution of the ‘ucld.mean’ parameter was set to an exponential distribution (mean=10.0, initial value=1.0). Four separate runs were performed in BEAST with 50 million generations each, sampling

parameters and trees every 1000 generations. Trees were summarized with burn-in values set to the first 25% of trees sampled using TreeAnnotator and were summarized in a maximum clade credibility tree.

1.3.6. Diversification analyses

In order to assess diversification patterns, we evaluated the state-dependent diversification of morphological characters using the Binary State Speciation and Extinction (BiSSE) and the MultiState Speciation and Extinction (MuSSE) models implemented in the R package ‘diversitree’ (FitzJohn, 2012), and the speciation/extinction and phenotypic/evolution models using BAMM v.1.0 (Rabosky, 2014). Four floral characters were chosen for these analyses because of their taxonomic interest: flower colors, flower size, spur length, and labellum position (Appendix 2). Three of these characters (flower color, flower size and spur length) have been used previously to delineate sections in Angraecum (e.g., Perrier de la Bathie, 1941; Garay, 1973) while the labellum position was added because it appeared useful for sectional delimitation. Since these methods require ultrametric and fully bifurcating trees, we used the BEAST tree as input.

Two characters were analyzed using BiSSE: flower color and labellum position. Since we were interested in the effect the green and white flower colors may have on diversification, the taxa that did not fit in these two colors (for instance, pink or purple flowered taxa) were excluded from the analyses using the ‘drop.tip2’ function of the R package ‘phyloch’ (Heibl, 2008). Three models were tested: (1) a full model that allows all parameters (λ: speciation rate, µ: extinction rate, q: transition rate) to vary, (2) a constrained model allowing speciation and transition rates to vary while keeping extinction rates equal between states (µ0 ~ µ1), and (3) a constrained model allowing extinction and transition rates to vary while keeping speciation rates equal between states (λ0 ~ λ1). Two characters were analyzed with MuSSE: flower size and spur length. Both characters were treated as ordered. A preliminary run was performed with the unconstrained model to determine the best transition model. For flower size, which has five character states (1: tiny, 2: small, 3: medium, 4: large, 5: very large), two linear transition models were tested: (1) a constrained model where character evolution was

possible in both directions between neighboring states (1 ↔ 2 ↔ 3 ↔ 4 ↔ 5), and (2) a constrained model where evolution could be either unidirectional or bidirectional between neighboring states (1 ←2 ←3 ↔ 4 → 5). For spur length, which has five character states (1: very short, 2: short, 3: medium, 4: long, 5: very long), two transition models were also tested: (1) a constrained linear model where character evolution was bidirectional and possible only among neighboring states (1 ↔ 2 ↔ 3 ↔ 4 ↔ 5), and (2) a constrained model where character evolution could be either unidirectional or bidirectional between states that are not necessarily adjacent (nonlinear) (1 ←2 ↔ 3 ↔ 4 ↔ 5, 1←3, 1←4). Having established the best fit transition model, we used it as a constraint while testing the three models of diversification: (1) an unconstrained model that allowed all parameters (λ, µ, q) to vary, (2) a constrained model allowing speciation and transition rates to vary while keeping extinction rates equal between states, and (3) a constrained model allowing extinction and transition rates to vary while keeping speciation rates equal between states. After testing the models with BiSSE and MuSSE, the posterior probabilities of the parameters were computed under a Bayesian framework after setting the priors to be exponential. The MCMC was run for 10000 generations sampling parameters every 100 generations, and the posterior distributions of parameters were summarized using the function ‘profiles.plot’ implemented in ‘diversitree’ (FitzJohn, 2012).

For the speciation/extinction model implemented in BAMM, we estimated our total sampling at 40% and we fractioned the data according to the number of species per genus being represented in our phylogeny (Appendix 4) in order to reduce the bias in estimating parameters under an assumption of incomplete sampling (Rabosky, 2014). To reduce the weight of the outgroup taxa we excluded them from the analyses using the ‘drop.tip2’ function of the R package ‘phyloch’ (Heibl, 2008). For the phenotypic/evolution model, we assessed the regime of morphological evolution of two of the four characters mentioned above, flower size and spur length (Appendix 4), since this model only treats continuous characters. For each model (speciation/extinction and phenotypic/evolution), BAMM was run for 5,000,000 generations and parameters were sampled every 1000 generations. The parameter priors were set using parameters generated from the R package ‘BAMMtools’ (Rabosky et al., 2014). The

rate shift configurations and the rate through-time generated from ‘bammdata outputs’ were analyzed using the R package ‘BAMMtools’.

Many criticisms have been made concerning the SSE’s family (BiSSE, MuSSE, GeoSSE, etc.) error rate and limitations (Maddison and FitzJohn, 2015; Rabosky and Goldberg, 2015). It has been reported that “within-clade pseudoreplications” might result in erroneously significant results (Maddison and FitzJohn, 2015). The use of additional methods like BAMM (Rabosky et al., 2013) has been proposed to reinforce the results implemented under SSE’s (Rabosky and Goldberg, 2015). Both methods use different approaches but are complementary: MuSSE looks at the effects of character states on diversification, while BAMM detects macroevolutionary rate shifts across phylogenetic trees.

1.3.7. Character state reconstruction

We examined the character evolution of four floral traits previously used in taxonomy (flower colors, flower size, spur length, and labellum position) using the Markov discrete character evolution (Pagel, 1994; Lewis, 2001) as implemented in the R package ‘diversitree’ (FitzJohn, 2012). Because this method requires ultrametric and fully bifurcating trees, we used the BEAST tree as input. No constraint was applied to the analyses leaving all parameters free.

1.4. Results

1.4.1. Phylogenetic relationships

Of the 69 newly sampled specimens that we sequenced, all were fully amplified with matK for a total length of 942 base pairs (bp); 66 with rps16 (1192 bp) except Angraecum sterophyllum, A. rhynchoglossum and Lemurella papillosa which failed to amplify; and 65 with trnL (1553 bp) except A. pseudofilicornu, A. rhynchoglossum, Oeoniella polystachys and Oeonia rosea. Only 13 of 69 samples were sequenced with ITS2 for a total length of 447 bp. When combined with GenBank sequences data, each individual matrix was composed of 190 taxa for matK, 140 for rps16, 170 for trnL and 88 for ITS2. Because of the small amount of

samples available with ITS2, we decided to not combine them with the plastid matrices. Therefore, the combined plastid matrix contained 3687 characters and 194 taxa. Sequences that were unavailable were treated as missing data. For the trnL intron, the ambiguous variable sites (360 bp) were excluded from our analyses. For the whole data set, 2361 characters were constant, 467 (13.8%) variable characters were parsimony-uninformative, and 537 (16%) characters were parsimony-informative. The combined plastid matrix produced five trees of 2202 steps with a consistency index (CI) of 0.56 and a retention index (RI) of 0.77.

The MrBayes, BEAST, and MP analyses were congruent. The 50% majority-rule consensus tree from Bayesian analyses of the combined plastid matrix is displayed in Fig. 1.1. Our results support a monophyletic subtribe Angraecinae (PP 1.0, BP 100). Two well- supported clades were identified within Angraecinae: clade I (PP 1.0, BP 77) comprised of Malagasy, IOI, African and American genera, and clade II (PP 1.0, BP 91) with Malagasy and IOI genera. Clade II had more branch support and showed more resolution than clade I. Two main subclades are observed within clade I: a Malagasy-IOI clade A (PP 1.0), and an African- American clade B (PP 1.0). Four Malagasy Angraecum s.l. species (A. perparvulum, A. cf. humile, A. pterophyllum, and A. rhynchoglossum) are nested within clade A. Three major subclades were observed in clade II: Aeranthes (PP 1.0, BP 98), Jumellea (PP 1.0, BP 97), and Angraecum (PP 1.0). Here, we define Angraecum as a monophyletic group including all Angraecum species and all other taxa nested within it in clade II. Based on branch support (PP, BP) and morphological resemblance, eleven clades are observed within Angraecum (Fig. 1.1, clade C to M). From the base to the top of the tree are: clade C (PP 1.0, BP 74) comprised members of sections Pectinaria, Pseudojumellea, Arachnangraecum and Filangis; clade D (PP 1.0, BP 81) comprising member of sections Perrierangraecum, Angraecum, Arachnangraecum and Filangis; clade E (PP 1.0) which includes A. sesquipedale, and A. sororium of section Angraecum; clade F (PP 1.0, BP 100) section Hadrangis; clade G (PP 1.0, BP 61) with section Humblotiangraecum and a member of section Perrierangraecum; clade H (PP 1.0, BP 94) section Boryangraecum; clade I (PP 1.0, BP 100) section Angraecoides; clade J section Arachnangraecum; clade K (PP 1.0, BP 98) section Angraecum (A. eburneum); clade L (PP 1.0, BP 75) comprised of members of sections Angraecum and Pseudojumellea; clade M (PP 1.0, BP 57) composed of sections Acaulia, Boryangraecum, Chlorangraecum,

Gomphocentrum, Lemurangis, and Lepervenchea. Clades E and J received weak support despite the fact that their positions were supported in a strict consensus tree. Two species are not included in any of the clades, A. nanum sister to clade I (PP 1.0, BP 72), and A. amplexicaule straddles between clade L and M (Fig. 1.1). Furthermore, two Malagasy genera, Oeoniella and Sobennikoffia, are nested within Angraecum. Even though many clades were strongly supported within Angraecum, resolution between clades was lacking.

The phylogeny obtained from ITS2 was lacking resolution and was slightly incongruent with the combined plastid ones (Appendix 5). Notably, the Malagasy Angraecinae genera Aeranthes, Jumellea, and Lemurorchis were embedded within Angraecum, rendering it paraphyletic. Nonetheless, all represented clades within Angraecum appeared to be congruent with those in the plastid data (Fig. 1.1). Our combined plastid and morphological analyses yielded a topology that is congruent with the combined plastid analyses with slight differences in branch supports (Appendix 6). Within Angraecum, clade M (PP 1.0, BP 75) gained support while clade D became weaker. The position of Lemurorchis was ambiguous, forming a polytomy with the Aeranthes – Jumellea clade in the molecular analyses while it was embedded with Angraecum in the combined analyses. The morphological analyses alone supported the monophyly of Angraecinae sensu Carlsward et al. (2006) with many polytomies observed within the clade (Appendix 7). According to the morphological phylogeny, most genera were monophyletic (Aeranthes, Jumellea, and most of the African genera) except Angraecum, where conflicts were also observed compared to the plastid tree.

Fig. 1.1. Phylogenetic relationships within subtribe Angraecinae. 50% Bayesian majority-rule consensus tree from combined plastid data (matK, rps16 and trnL). Values above branches or at nodes represent posterior probability (PP) and bootstrap percentage (BP) support. Dashes represent branches that collapsed in MP strict consensus tree. Taxa with asterisk are Angraecum sensu Garay species. Abbreviations in brackets denote sections sensu Garay (1973): Aca = Acaulia, Agd = Angraecoides, Ang = Angraecum, Arc = Arachnangraecum, Bor = Boryangraecum, Chl = Chlorangraecum, Fil = Filangis, Gom = Gomphocentrum, Had = Hadrangis, Hum = Humblotiangraecum, Lem = Lemurangis, Lep = Lepervenchea, Pct = Pectinaria, Per = Perrierangraecum, Psj = Pseudojumellea.

Angraecum tenuispica (Lep) Angraecum pauciramosum (Lep) 1/80 Angraecum musculiferum (Lep) Angraecum cf.pauciramosum (Lep) 1/84 Angraecum caricifolium (Lep) 1/68 Angraecum madagascariense (Lem) Angraecum tenuifolium (Lep) Angraecum obversifolium (Agd) 1/98 Angraecum cf.sacculatum (Gom) 1/63 Angraecum costatum (Lem) 1/61 1/100 Angraecum pinifolium (Bor) Angraecum cf.pinifolium (Bor) Angraecum sacciferum (Bor) clade M

1/69 Angraecum hermannii (Pct) Angraecum cordemoyi (Gom) Angraecum baronii (Lem) Angraecum ochraceum (Aca) 1/93 Angraecum chaetopodium (Aca) 1/57 1/78 Angraecum setipes (Aca)

1/93 Angraecum cf.zeylanicum (Gom) Angraecum caulescens (Gom) 0.96/- 1/93 Angraecum huntleyoides (Chl) Angraecum calceolus (Gom) 0.9/- Angraecum multiflorum (Gom) Angraecum amplexicaule (Psj) 0.92/- Angraecum serpens (Ang) 1/74 Angraecum moratii (Fil) 1/90 Angraecum florulentum (Ang) 1/73 clade L 1/75 Angraecum penzigianum (Ang) Angraecum oblongifolium (Per) Angraecum longicalcar (Ang) 1/68 Angraecum eburneum var.superbum (Ang) 1/98 clade K Angraecum eburneum var.xerophilum (Ang) Angraecum eburneum var.eburneum (Ang) Oeoniella polystachys 1/- 1/89 Angraecum expansum (Arc) Angraecum conchiferum (Arc) Angraecum linearifolium (Arc) 1/71 Angraecum corrugatum (Arc) 1/74 1/98 Angraecum conchoglossum (Arc) Angraecum mirabile (Arc) clade J 1/98 Angraecum cf.germinyanum (Arc) 1/99 Angraecum germinyanum (Arc) 1/90 Angraecum arachnites (Arc) Angraecum appendiculatum (Arc) Angraecum scottianum (Arc) 1/72 Angraecum pseudofilicornu (Arc) Angraecum triangulifolium (Agd) 1/100 Angraecum sedifolium (Agd) clade I 1/71 Angraecum pingue (Agd) Angraecum nanum (Nana) 1/94 Angraecum teres (Bor) clade H Angraecum dives (Bor) Angraecum leonis (Hum) 0.9/- 1/97 Angraecum compactum (Per) Angraecum clareae (Per) clade G 1/61 Angraecum viguieri (Hum) 1/100 Angraecum magdalenae (Hum)

Sobennikoffia humbertiana Angraecum striatum (Had) 1/- 1/100 Angraecum bracteosum (Had) clade F Angraecum cadetii (Had) Angraecum sororium (Ang) 1/- clade E Angraecum sesquipedale (Ang)

Angraecum pseudodidieri (Per) Angraecum 1/- Angraecum elephantinum (Per) Angraecum didieri (Per) Angraecum cf.rutenbergianum (Per) Angraecum ankeranense (Per) Angraecum liliodorum (Per) 1/65 Angraecum cf.liliodorum (Per) Angraecum obesum (Per) Angraecum peyrotii (Per) 1/86 Angraecum cf.peyrotii (Per) 1/97 Angraecum cf.breve (Per) 1/59 clade D Angraecum borbonicum (Per) Angraecum sterophyllum (Per) Angraecum cf.elephantinum (Per) 1/97 Angraecum letouzeyi (Per) 1/60 Angraecum cornigerum (Fil) Angraecum bicallosum (Per)

1/81 Angraecum drouhardii (Per) 1/99 Angraecum cucullatum (Per) Angraecum lecomtei (Per) Angraecum praestans (Hum)

Angraecum ramosum (Arc)

Angraecum dasycarpum (Pct)

Angraecum pectinatum (Pct) 1/57 Angraecum mauritianum (Psj)

Angraecum danguyanum (Arc) 1/64 clade C Angraecum cf.panicifolium (Pct) 1/74 Angraecum filicornu (Fil)

Angraecum panicifolium (Pct) 1/95 Angraecum cf.humblotianum (Pct)

Jumellea tenuibracteata clade II 1/91 Jumellea francoisii

Jumellea bosseri 1/77 Jumellea arborescens

Jumellea rigida

Jumellea maxillarioides

Jumellea exilis

Jumellea spathulata 1/99 Jumellea longivaginans Jumellea

Jumellea hyalina 1/97 1/54 Jumellea densifoliata 1/100 Jumellea arachnantha

Jumellea jumelleana

Jumellea majalis 1/95 Jumellea brevifolia

Jumellea anjouanensis

Jumellea teretifolia

Aeranthes virginalis 1/54 1/87 Aeranthes grandiflora

Aeranthes arachnites

1/66 Aeranthes strangulata Aeranthes neoperrieri Aeranthes 1/100 Aeranthes longipes

Aeranthes caudata 1/98 Aeranthes moratii 1/98 Aeranthes aemula

Aeranthes tenella

Lemurorchis madagascariensis

Pectinariella pungens. 1/72 Pectinariella gabonense. 1/96 Pectinariella atlantica. 1/96 Pectinariella doratophylla. Pectinariella subulata. Ypsilopusviridiflorus Ypsilopuslongifolius Rangaeris amaniensis Rangaeris muscicola 1/83 Tridactyle filifolia Tridactyle bicaudata Podangis dactyloceras

1/99 Cyrtorchis chailluana ANGRAECINAE Cyrtorchis arcuata 1/100 Listrostachys pertusa 1/86 Aerangis hildebrandtii Aerangis hariotiana 1/67 Aerangis punctata Aerangis macrocentra Aerangis ellisii 1/98 Dolabrifolia disticha. 1/96 Dolabrifolia bancoense. 1/99 Dolabrifolia cf.aporoides. Dolabrifolia podochiloides.

1/91 Eurychone rothschildiana Eurychone galeandrae 1/100 Bolusiella maudiae Bolusiella iridifolia clade B 1/100 Cribbia confusa 1/61 Cribbia brachyceras 1/64 Rhipidoglossumsubsimplex Rhipidoglossumxanthopollinium 1/61 1/100 Mystacidium flanaganii Mystacidium capense 0.94/52 Sphyrarhynchus schliebenii Angraecopsis parviflora

1/100 Solenangis wakefieldii Solenangis clavata 0.9/- 1/100 Microcoelia stolzii Microcoelia gilpinae

1/96 Ancistrorhynchus recurvus Ancistrorhynchus capitatus Calyptrochilumchristyanum 1/94 Angraecoides erecta. 1/99 Angraecoides cultriforme. 1/73 Angraecoides chevalieri. 1/96 Angraecoides cf.moandense . 1/96 Dendrophylax funalis 1/68 Campylocentrumsullivanii Eichlerangraecuminfundibulare. 1/100 clade I 1/77 Eichlerangraecumcf.eichlerianum. Eichlerangraecumbirrimense. 1/84 Diaphananthe vesicata 1/100 Diaphananthe sarcophylla Diaphananthe pellucida 1/80 Cryptopus paniculatus 1/80 Cryptopus elatus

0.9/- Oeonia rosea Neobathiea grandidierana 0.9/- 1/98 Lemurella papillosa Lemurella pallidiflora 0.94/- Angraecumpterophyllum. clade A 1/78 Angraecum cf.humile. 0.95/51 0.95/- Angraecumperparvulum. Beclardia macrostachya Angraecumrhynchoglossum. Erasanthe henricii Vanda tricolor 1/93 Acampe ochracea 1/100 Aerides odorata outgroup Phalaenopsis cornu-cervi Polystachya fulvilabia

1.4.2. Divergence time estimates

The maximum credibility tree of the calibrated relaxed molecular clock analysis of Angraecinae is shown in Appendix 8. Our results suggest that Angraecinae shared a most recent common ancestor (MRCA) in the late Oligocene (~ 26.1 Ma), and started to diverge in the early Miocene. Diversification started at approximately 21.1 Ma (giving a node height highest posterior density (HPD) intervals at 95% ranging from 18.2 – 33.5 Ma) for clade I and 17.18 Ma (95% HPD: 14.6 – 27.8 Ma) for clade II. In clade I, the African-American clade B diverged at approximately 19.46 Ma (95% HPD: 14.6 – 27.8 Ma), while the Malagasy-IOI clade A diverged at approximately 18.1 Ma (95% HPD: 14.6 – 27.8 Ma). Within clade II, Angraecum, Aeranthes and Jumellea started to diversify at around 14.56 Ma (95% HPD: 9.9 – 19.2 Ma), 10.7 Ma (95% HPD: 5.8 – 15.2 Ma), and 7.8 Ma (95% HPD: 4.9 – 11 Ma) respectively. Two waves of diversification are observed in Angraecum, during the Pliocene (~ 6 – 2.6 Ma) and during the Pleistocene (~ 2.6 – 0.2 Ma). Angraecum section Hadrangis which is endemic to the Mascarene Islands diverged at approximately 1.66 Ma (95% HPD: 0.4 – 3.3 Ma), and the divergence time for the two species A. bracteosum and A. striatum that are endemic to Reunion is estimated at 0.2 Ma (95% HPD: 0 – 0.7 Ma). Since this age is younger than the estimated age of the Island, approximately 2.1 Ma (Baksi and Hoffman, 2000), it is reasonably consistent with the age obtained in our analyses.

1.4.3. Species diversification

Results from BiSSE showed that the second model (µ0 ~ µ1) received the best AIC score for flower color and labellum position (Table 1.1). The green and white colors had equal rates of speciation (Fig. 1.2A), while the uppermost labellum showed a higher speciation rate compared to the lowermost one (Fig. 1.2B). Results from MuSSE showed that the second transition model had the best AIC score (Table 1.1); this model was used to test the diversification models. The diversification model with extinction rates equal between states

Table 1.1. Comparaison of diversificaiton model used for the BiSSE and MuSSE analyses. Bold indicates the best fit model selected to test diversification of Angraecinae. Arrow represents transition allowed between states: ↔, reversible, →, irreversible. Parameters: λ, speciation rate; µ, extinction rate; q, character transition rate. Character states: A0, green; A1, white; B0, uppermost; B1, lowermost; C1, tiny; C2, small; C3, medium; C4, large; C5, very large; D1, very short; D2, short; D3, medium; D4, long; D5, very long. Abbreviations: AIC, Akaike information criterion; M, model tested; lnLik, Log likelihood.

Character Model lnLik AIC A color of flower full -567.496 1146.992 M1 (λ0 ~ λ1) -567.604 1145.208 M2 (µ0 ~ µ1) -567.496 1144.992 M3 (q01 ~ q10) -569.488 1148.976 B position of labellum full -599.017 1210.035 M1 (λ0 ~ λ1) -601.055 1212.110 M2 (µ0 ~ µ1) -599.017 1208.035 M3 (q01 ~ q10) -599.018 1208.036 C size of flower (pre-run) full -716.052 1492.105 M1 (1 ↔ 2 ↔ 3 ↔ 4 ↔ 5) -723.425 1482.851 M2 (1 ←2 ←3 ↔ 4 → 5) -722.499 1474.999 size of flower full -722.499 1474.999 M1 (λi ~ λj) -722.666 1467.332 M2 (µi ~ µj) -726.511 1475.022 D spur length (pre-run) full -727.404 1514.808 M1 (1 ↔ 2 ↔ 3 ↔ 4 ↔ 5) -737.145 1510.290 M2 (1 ←2 ↔ 3 ↔ 4 ↔ 5, 1←3, 1←4) -729.235 1496.471 spur length full -729.235 1496.471 M1 (λi ~ λj) -729.235 1488.471 M2 (µi ~ µj) -735.279 1500.559

received the best score for flower size and spur length, suggesting that character states have an effect on speciation. Medium and large flowers showed higher speciation rates compared to tiny and small flowers (Fig. 1.2C). All spur length states had a similar effect on speciation rates (Fig. 1.2D).

Fig. 1.2. Posterior probability distributions for the speciation rates (in Ma) of morphological characters using equal rate speciation (µ0 ~ µ1) with the BiSSE model and equal rate extinction (λi ~ λj) with the MuSSE model: flower color (A), labellum position (B), flower size (C), and spur length (D). Abbreviation: v, very.

The speciation/extinction model from BAMM revealed 5 distinct configuration shifts from the 95% credible set (Fig. 1.3) of which 66% of the samples in the posterior distribution showed no shift, 14% showed a single shift at the node of clade II, 6.6% had one within the Aeranthes clade, 5.9% had one shift at the branch of Beclardia macrostachya, and 3.6% of the posterior distribution had one shift at the branch of clade I. Our BAMM results showed a

decline of speciation rate-through-time (RTT) for Angraecinae, starting from approximately 0.39 during the Miocene to 0.23 towards the present (Fig. 1.3).

Fig. 1.3. Configuration shifts from the 95% credible set sampled by BAMM from the Angraecinae phylogeny and evolutionary rates through time. The intensity of colors on branches reflects the relative probability density of speciation rates (cool colors = slow, warm = fast). Black circles denote the position of the macroevolutionary regime shifts present in each sample. Blue curve indicates the mean speciation rate-through-time trajectory of Angraecinae in million years.

The phenotypic/evolutionary model showed 32 distinct configurations on spur length. The five that received the best sample frequencies are displayed in Fig. 1.4. Four main shifts are observed, one at the branch of Eichlerangraecum clade, one at the branch of clade A, one at the branch of Angraecum appendiculatum and one shift at the branch of Angraecum corrugatum. The RTT phenotypic evolution showed an increased rate in spur length starting from 0.02 during the Pliocene to 0.08 in the Pleistocene and to the present (Fig. 1.4). No shift has been detected regarding flower size within Angraecinae and the RTT phenotypic evolution was constant (Appendix 9).

1.4.4. Ancestral state reconstructions

Our results showed that a labellum in the lower position is plesiomorphic in Angraecinae, while an upper labellum is apomorphic and evolved at least five times independently (Fig. 1.5). White flowers appear to be symplesiomorphic in Angraecinae, while green flowers are apomorphic and evolved several times in Angraecinae and two times independently in Angraecum (Appendix 10A). A medium flower appears to be the ancestral state in Angraecinae, while large and small flowers are derived (Appendix 10B). Long spur is the ancestral state and short spur is derived and arose several times independently (Appendix 10C). Our results showed that the color, the size, and the spur length are homoplastic within Angraecum. The taxa that showed uppermost labella appeared to be monophyletic and received very strong support in the phylogeny (Angraecoides, Dolabrifolia, Pectinariella, clades H to M).

Fig. 1.5. Ancestral state reconstructions of floral traits in Angraecinae implemented in ‘diversitree’; colors represent character states and pie charts represent the probability of ancestral states at nodes. Asterisks (*) indicate taxa illustrated in the pictures to the right to represent the flower shape of each clade except for Jumellea which is represented by Jumellea comorensis (not sampled in the phylogeny). Photo: Andriananjamanantsoa.

1.5. Discussion

1.5.1. Systematics of Angraecinae

The problem we encountered when amplifying the ITS region was primarily caused by endophytes: instead of amplifying orchid DNA, the primers amplified fungi. This was also reported by several authors while working with the Orchidaceae (e.g. Carlsward et al., 2006). The incongruence observed with the morphological phylogeny illustrates the existence of homoplasies in Angraecinae and the difficulty of delineating natural groups using morphology in this group. The variation in branch support between the combined plastid and combined plastid + morphological analyses are probably due to synapomorphic characters that were observed in some clades but were lacking in others, as well as strong plesiomorphic characters that are encountered in subtribe Angraecinae. The poor resolutions observed within or between clades were probably due in part to a lack of sampling or caused by low genetic variation. More regions are needed to bring more resolution to the phylogeny, especially on the relationships between clades.

Our results confirmed the monophyly of Angraecinae and the distinction of the Malagasy–IOI clades from the African–American one (Carlsward et al., 2006, Micheneau et al., 2008a). Our results confirmed also the polyphyly of Angraecum s.l. Indeed, apart from the African sections already transferred to new genera (Szlachetko et al., 2013) four additional Malagasy species were embedded with clade A: Angraecum cf. humile, A. perparvulum, A. pterophyllum, and A. rhynchoglossum belonging to sections Lemurangis, Nana, Pectinaria, and section Acaulia respectively (Garay, 1973). Micheneau et al. (2008) first reported the existence of rare Angraecum species from Madagascar and the Mascarenes (A. sp. TP84, section Nana) in clade A. In our treatment, this taxon was unrelated to ours, and embedded with Lemurella (result not shown). To keep Angraecum monophyletic, these species need to be removed from the genus.

1.5.2. Systematics of Angraecum

The paraphyly observed in the ITS2 data could be due to a lack of sampling and of low genetic variation. Carlsward et al. (2006) had detected a paralogy problem in the ITS region in the Malagasy Angraecinae and most notably in clade II and therefore they excluded this region from subsequent analyses, including the genera Aeranthes, Jumellea, Lemurorchis, Oeoniella and Sobennikoffia. Since most of the sequences we used to reconstruct the ITS phylogeny came from their work, except for Angraecum which came mostly from our samples, and given that only ITS2 was sequenced by us, we believe that it is difficult to reconciliate the data at this point. Given the potential variability of ITS sequences, it might be worth attempting to resolve the problems of paralogy and primer design in order to increase our knowledge of Angraecinae.

The plastid results also showed that Angraecum is paraphyletic. Two Malagasy genera, Oeoniella and Sobennikoffia, are nested within this clade, as was reported by Carlsward et al. (2006). However, Aeranthes and Jumellea form natural groups sister to Angraecum as also reported by Micheneau et al. (2008). Schlechter (1918) admitted the close resemblance between Oeoniella and Angraecum, but emphasized the differences in column shape and stipe length. These two genera differ from Angraecum in their labellum shape. If we look closer at Oeoniella and unfold the labellum, the flower looks similar to Angraecum eburneum, except that the perianth is soft, the labellum spurless, and the ovary untwisted. Sobennikoffia, a small genus of four species, was included within Angraecum until transferred to a new genus by Schlechter (1925) because of the three-lobed labellum.

Of the 16 sections of Garay (1973) represented in our phylogeny, only one is monophyletic, the Mascarene section Hadrangis, all other sections are paraphyletic (Gomphocentrum, Lemurangis and Lepervenchea) or polyphyletic (Acaulia, Angraecum, Angraecoides, Arachnangraecum, Boryangraecum, Filangis, Humblotiangraecum, Pectinaria, Perrierangraecum and Pseudojumellea). Micheneau et al. (2008a) mentioned the unnaturalness of Garay’s sections, and pointed out the complexity of dealing with the sections with small and greenish flowers. We decided to not consider A. nanum within clade I because of their morphological differences. Although these species produce green flowers, A. nanum which represents section Nana, is characterized by tiny plant with racemose inflorescences

and minute flowers, while clade I is characterized by erect or pendent plants with one- flowered inflorescences and larger flowers. The position of A. amplexicaule is ambiguous but in the BEAST analyses it was embended within clade L (Appendix 8). Morphologically, A. amplexicaule has an inflorescence and flower shape similar to those of species in clade L except that the habit is more robust within clade L and the leaves are coriaceous. Descpite the weak support, we decided to include A. pseudofilicornu and A. scottianum in clade J because all species in this clade share the same inflorescence type and flower shape except that the sepals and petals are reduced and the habit crassulescent with the two species. Furthermore, in the ITS2 topology A. scottianum was embended within the clade (Appendix 5).

The character (type of inflorescence, flower size and color, and spur length) used by many authors (e.g., Schlechter, 1925; Perrier de la Bathie, 1941; Garay, 1973; Stewart et al., 2006; Cribb and Hermans, 2009) to delineate sections in Angraecum are of limited taxonomic interest and their distribution is not coherent with the phylogeny. Spur length and flower size are generally correlated (small flowers have short spur and large flowers have long ones), with the exception of A. appendiculatum and A. corrugatum. These two Mascarene species, often considered peloric forms of A. arachnites and A. conchoglossum, respectively (Garay, 1973; Hermans and Cribb, 2005; Micheneau et al., 2008a), instead lost or have changed the genes responsible for labellum and spur development (Hossain and Levy, 2014). The phenotypic plasticity observed in these two species was considered to be the result of species radiation (Micheneau et al., 2008a). But colonization of new habitats (with new selection regimes) might affect gene expression that is responsible for floral development (Theissen and Saedler, 1995; Hsu et al., 2003; Chang et al., 2010; Pan et al., 2011; Ding et al., 2013). Chang et al. (2010) showed that the size and shape of sepal/petal/labellum in Oncidium are regulated by the OMADS5 gene. These species are spurless, the labellum changed from suborbicular concave, typical in section Arachnangraecum, to linear-lanceolate, and they became self-pollinated (Pailler et al., 2013). There is no evidence here of a loss of bilateral symmetry (there is still slight zygomorphy) that would be expected in peloric flowers. Alternately, variation in spur length could be due to genotypic drift (Mallet et al., 2014; Luo et al., 2015). For instance, Stewart et al. (2006) observed that spur length in subspecies of A. eburneum was shorter the further away from Madagascar a subspecies was.

1.5.3. Labellum position a missing character to delineate sections in Angraecum

After evaluating several morphological characters (Appendix 2), we noted that the position of the labellum, uppermost or lowermost, coincided with the delimitation with clades in Angraecum (Fig. 1.5). The degree of resupination in an orchid flower can vary from 0° to approximately 360° depending on the inflorescence (Arditti, 2003), and might be specific to a taxon. The 180° resupination results in a lowermost labellum and a 360° resupination (double twist of pedicel or ovary) in an uppermost labellum. The position of the labellum has been used in Orchidaceae to delimit sections within genus Bulbophyllum (Fischer et al., 2007), and in Gesneriaceae to delimit the genus Alloplectus (Clark and Zimmer, 2003). The newly defined African genera (Szlachetko et al., 2013) Angraecoides, Dolabrifolia, Pectinariella are also characterized by an uppermost labellum. Our results showed that Clades C to G have lowermost labella, while clades H to M are composed exclusively of species with uppermost labella. The dispersion of the species of section Angraecum sensu Garay (Fig. 1.1: clade E vs clade K) is a concrete example demonstrating the usefulness of this character to delimit sections. The labellum position is difficult to observe on herbarium sheets and often the only way to clearly see it is on living specimens. This could explain the fact that none of the original species descriptions mentioned this character (Schlechter, 1925; Perrier de la Bathie, 1941; Garay, 1973). Given these new findings, species description should be updated.

1.5.4. Temporal framework and paleoclimate events in Angraecum

Our age estimate for Angraecinae and Angraecum is approximately 4 Ma older than that obtained by Micheneau et al. (2010); the age of two endemic species of the Reunion Island (A. bracteosum and A. striatum) is younger in our results, however. This can be explained by our calibration. We set the root of the tree to 35 Ma following Gustafsson et al. (2010), while Micheneau et al. (2010) followed Ramirez et al. (2007) and fixed it at 30.37 Ma. The main difference between the two calibrations is based on the number of fossils used in the analyses: Ramirez et al. (2007) used a single fossil, whereas Gustafsson et al. (2010) included three to calibrate their phylogeny of the Orchidaceae. We did not use secondary calibrations,

while Micheneau et al. (2010) used island ages as constraints in their analysis. Since no fossil was available to directly calibrate our phylogeny, these age estimates remain approximations. Nonetheless, our calibration is consistent with the divergence time of the Vandeae estimated by Givnish et al. (2015), and the divergence time estimate for Angraecum matches well with the diversification age of most Malagasy Angiosperm endemic genera (Buerki et al., 2013). During the Miocene, the climate of Madagascar shifted gradually from cool dry to warm humid. Contraction of the arid forest and expansion of the current tropical forest in Madagascar have been documented as a result of the northern migration of the island towards the equator (Buerki et al., 2013). This migration was associated with the establishment of the trade winds, and later of the monsoons, which increased moisture levels throughout eastern Madagascar. Two diversification events were observed in Angraecum, during the Pliocene and in the Pleistocene. This could be explained by the Quaternary glaciations event, where the climate fluctuated between cold to warm (Ruddiman, 2001). The diversity increases recorded within Aeranthes, Angraecum and Jumellea during the Pleistocene accounts for approximately 60% of their species (Appendix 8). This coincides with the radiation bursts observed in Malagasy tree ferns during the same period (Janssen et al., 2008) which somehow reflect the presence of humid and warm climate.

1.5.5. Diversification in Angraecum

The BiSSE and MuSSE results revealed that flower size and labellum position appear associated with the diversification of Angraecinae and Angraecum. An uppermost labellum concurred with a higher speciation rate compared to a lowermost labellum. Medium and large flowers were associated with higher speciation rates when compared to minute, small or very large flowers. The overlap in the posterior probabilities observed in flower color and spur length (Fig. 1.2) could be interpreted as an equal rate of speciation between color and length categories. Floral divergence in Orchidaceae has been associated with pollinator shifts (Peter and Johnson, 2014). Fischer et al. (2007) pointed out the importance of flower orientation in plant evolution and on species diversification. Lowermost labella serve as landing platforms for pollinators (Fischer et al., 2007), while uppermost labella are associated with either

autogamy or a switch to pollinating insects that prefer walking rather than flying (Rudall and Bateman, 2002). Long spurs have been shown to be associated with specific pollinators in Angraecum (Nilsson et al., 1987; Wasserthal, 1997). Because of this specificity, long-spurred flowers are more efficiently pollinated (Fulton and Hodges, 1999; Hodges et al., 2004; Boberg et al., 2014), though it does not necessarily result in a greater speciation rate, as appears to be the case in Angraecum.

The speciation/extinction analyses obtained from BAMM model could be resumed to three best configuration shifts in Angraecinae, one shift at the MRCA of clade I, one at the MRCA of clade II, and one within the Aeranthes clade. BAMM detected general shifts where diversification could potentially have arisen (at the MRCA of Aeranthes, Angraecum, and Jumellea), but could not detect evidence on specific clades associated with diversification regimes that we expected within Angraecum (the shift from lowermost to uppermost labella, or the high speciation rate detected by BiSSE with flower color). A lack of performance of the statistical models used in BAMM was pointed out by Rabosky and Goldberg (2015). The shift detected within Aeranthes could be a bias of incomplete sampling (Fig. 1.3). Within Jumellea we selectively sampled at least one representative of each section of the genus (Rakotoarivelo et al., 2013), while sampling was more random in Aeranthes. Better samples are required before any conclusion on the diversification regimes operating in this clade can be drawn.

The phenotypic/evolution analyses obtained from BAMM model showed that there was no shift associated with labellum size, while at least four shifts were detected with spur length (Fig. 1.4). This leads us to conclusion that flower size is evolutionary constant within clades in the whole subtribe Angraecinae, while the evolution of spur length is more variable in some clades than in others. Rakotoarivelo et al. (2013) pointed out the lability of spur length in Jumellea, and its inefficiency on delimiting sections. Since flower size is generally conserved within clades, it is difficult to ascertain the extent to which this character may influence speciation, even though this character usually is associated with pollinator type. The shifts observed in clade C and J with spur length evolution suggest that these clades are the most phenotypically diverse in the whole Angraecinae. However it does not necessary mean that there is species diversification associated with the rate of spur length evolution. If we look at clade C, there is no phylogenetic signal showing that short or long spurs led to more

speciation, which excludes all hypotheses of high rate diversification associated with this character. The rate shifts observed in clade J were caused by the two Mascarene species A. appendiculatum, and A. corrugatum.

Our BAMM analyses showed a slowdown in time of the diversification rate in Angraecinae in general (Fig. 1.3). The results showed an accumulation of lineages rather than rapid radiation in Angraecum. To test for homogeneity in the diversification rate through time, we calculated the gamma statistic of Pybus and Harvey (2000) as implemented in the R package ‘ape’. Our significantly negative result (γ = – 2.204816) rejects the null hypothesis for a constant-rate, suggesting that the speciation rate was initially high but slowed through time, which is congruent with the BAMM results. This early burst in lineages could be explained by the fact that Angraecum species develop a crassulacean acid metabolism (CAM) which allows them to tolerate desiccation (Kluge et al., 1997). Whitman et al. (2011) reported that lithophytic Angraecinae are tolerant to limited moisture availability, and chronic bushfire did not kill the population of Angraecum sororium and Jumellea rigida but only reduced their expansion. But, diversification slowdown could be also the result of ecological limitation, competition, access to pollinators, species carrying capacity (Gavrilets et Vose, 2005; Phillimore et Price, 2008; Jonsson et al., 2012), or maintenance of niche similarity (Moen et Morlon, 2014). Lineage diversiﬁcation is usually high right after colonization of a new niche, but slows through time as niches get occupied and ecological conditions for speciation decrease (Reddy et al., 2012). Scantlebury et al. (2013) reported that diversification slowdown in Malagasy fauna is a general pattern of adaptive radiation like that observed in amphibians and birds (Vences et al., 2009; Jonsson et al., 2012; Reddy et al., 2012). However, it has been shown that the diversification pattern of the tropical rain forest has been an accumulation of lineages through time and not sudden adaptive radiations (Couvreur et al., 2011). Our results appear to support this hypothesis.

It is intriguing that each clade in Angraecum is associated with specific floral traits. Aeranthes and Jumellea are species rich and occupy the same ecological niche as Angraecum. The three genera diverged approximately at the same time, but Angraecum has more species. If we look at the morphological differentiation between the three genera, Jumellea and Aeranthes have distinctive traits that characterize them as clades, while Angraecum has

variable traits specific to each subclade. The diversity of Angraecum could perhaps be partly explained by its morphological evolution. Taxonomic diversity has been demonstrated to be complemented by morphological and ecological diversity (Shi and Rabosky, 2015). Pollinators could have played a significant role in this diversification process. Little is known about Angraecum pollinators, but we know now that it is pollinated by different kinds of pollinators (Micheneau et al., 2006; Micheneau et al., 2008c; Micheneau et al., 2009; Pailler et al., 2013) not only hawk-moths as traditionally thought (Nilsson et al., 1985; Nilsson et al., 1987; Wasserthal, 1997). Micheneau et al. (2008a) reported that the radiation of Angraecum in the Mascarenes was caused by a change in pollinators on these Islands. The absence of the original pollinators is considered to have resulted in auto-pollination in some Angraecinae species (Micheneau et al., 2008b; Hermans and Hermans, 2013; Pailler et al., 2013), which could be associated with a decrease or loss of rewards such as fragrance or nectar for the species (Pailler et al., 2013).

1.6. Conclusion

The present study presents the most comprehensive phylogenetic reconstruction of genus Angraecum to date including all the sections sensu Garay except section Afrangraecum. With the African sections being removed (Szlachetko et al., 2013) Angraecum counts currently ca. 190 species (including varieties). Our results confirmed the paraphyly or polyphyly of Angraecum sections, but showed the morphology to be consistent with the phylogeny. The position of the labellum, lowermost or uppermost, allowed us to delineate several clades. An updated systematic revision of the genus is required considering these findings. Our study revealed that many characters are associated with species diversification of Angraecum, the orientation of the labellum being one. However, our analyses failed to detect shifts that could have been caused by morphological diversification. Overall, the evolution and diversification of Angraecum resulted from species accumulation through time rather than rapid radiation.

1.7. Acknowledgments

The authors thank A. Schuiteman (Royal Botanic Gardens, Kew) and M. Pignal (Museum of Paris) for their help in accessing herbarium specimens. We would like to thank the University of Antananarivo and Madagascar National Parks for granting research permits. We are thankful to M. Kuzmina and the Canadian Center for DNA Barcoding for help with sequencing and the Jodrell laboratory for providing genetic material for this study. We are very thankful to C. Micheneau, R. Lei, S. Joly, J. Saarela, F.J. Lapointe, E. Gagnon, M. Light, and S. Renaut for their assistance, comments and expertise. Computations were made on the supercomputer “Cottos” and “Briaree” from the University of Montreal, managed by Calcul Québec and Compute Canada. The operation of this supercomputer is funded by the Canada Foundation for Innovation, Ministère de l'Économie, de l'Innovation et des Exportations du Québec and the Fonds de recherche du Québec - Nature et technologies. This study was supported by Omaha’s Henry Doorly Zoo and Aquarium, and the Madagascar Biodiversity Partnership.

Chapitre 2: Systematics and taxonomic revision of genus Angraecum (Orchidaceae, Angraecinae) and two new genera in the orchids of Madagascar

Herinandrianina N. Andriananjamanantsoa and Luc Brouillet

Département de sciences biologiques, Université de Montréal, 90 Av. Vincent d’Indy, Montréal (QC) H2V 2S9, Canada

Journal ciblé: Phytotaxa (Modifié pour les besoins de la présente thèse)

Tous les auteurs ont donné leur autorisation pour ce manuscrit à être inclus dans cette thèse

Contribution respective des auteurs:

Tahiana Andriananjamanantsoa: Conception du projet, collecte, traitement et analyse des données, interprétation et rédaction

Luc Brouillet: Supervision, conseils, recommandations, corrections et direction générale

2.1. Résumé/ Abstract

Résumé

Angraecum est le plus grand genre de la sous-tribu des Angraecinae avec ca. 195 espèces. Les études phylogénétiques moléculaires récentes appuyées par des données morphologiques ont montré qu’Angraecum doit être recirconscrit afin d'en préserver la monophylie. Un synopsis du genre Angraecum avec les 14 sections que nous reconnaissons ici est présenté, incluant cinq sections nouvelles: Africanae, Oeoniella, Robusta, Sobennikoffia et Stellariangraecum. Deux nouveaux genres, Acaulia et Parangraecum, sont proposés pour accommoder des espèces phylogénétiquement séparées du genre Angraecum.

Mots-clés: Acaulia, Angraecinae, Angraecum, Orchidaceae, Parangraecum, classification des sections

Abstract

Angraecum is the largest genus of subtribe Angraecinae with ca. 195 species. Recent molecular phylogenetic studies supported by morphological data showed that Angraecum must be re-circumscribed to maintain monophyly. A synopsis of Angraecum with 14 sections recognized here is presented, including five new sections: Africanae, Oeoniella, Robusta, Sobennikoffia and Stellariangraecum. Two new genera, Acaulia and Parangraecum, are proposed to accommodate species phylogenetically segregated from Angraecum.

Keywords: Acaulia, Angraecinae, Angraecum, Orchidaceae, Parangraecum, sectional classification

2.2. Introduction

Angraecum is the largest genus in subtribe Angraecinae (Orchidaceae, tribe Vandeae) and includes about 195 species (chap. 1). Recent molecular and morphologic phylogenetic studies have shown that Angraecum is polyphyletic as traditionally circumscribed and that the sections proposed by Garay (1973) are often poly- or paraphyletic (Micheneau et al., 2008a; chap. 1). To maintain monophyly of the genus and sections, new circumscriptions are required. Traditionally, sectional classification has been based on morphological characters, especially floral traits (e.g., Garay, 1973; Stewart et al., 2006; Cribb and Hermans, 2009). These characters have been used as well to recognize informal groups within Malagasy Angraecinae like Jumellea (Perrier de la Bathie, 1941), but have been demonstrated to be evolutionary labile (Rakotoarivelo et al., 2012; chap. 1) and of limited use for systematic purpose. In order to render Angraecum monophyletic, we propose to include all species that are nested within the Angraecum clade within a single genus, Angraecum, as suggested by Andriananjamanantsoa et al. (chap. 1), and to remove elements that are found within other clades of subtribe Angraecinae. Likewise, we are assigning to sections all species grouped within a clade (Fig. 2.1). Four species were removed from Angraecum (A. cf. humile, A. perparvulum, A. pterophyllum, and A. rhynchoglossum) and accommodated into two new genera (Acaulia and Parangraecum).

Fig. 2.1. Partial phylogenetic relationships of Angraecum showing ancestral state reconstructions of labellum position implemented in ‘diversitree’ (chap. 1); complete tree presented in upper left corner, with selected partial trees in red-dashed boxes shown on the right; taxa in bold are the new genera Acaulia and Parangraecum. Colors represent labellum character states (orange, uppermost; green, lowermost) and pie charts represent the probability of ancestral states at node. Abbreviation: A, Angraecum.

A.cf.pauciramosum A.tenuispica A.musculiferum A.pauciramosum A.caricifolium A.madagascariense A.tenuifolium A.cordemoyi A.cf.sacculatum A.obversifolium A.costatum A.hermanni A.sacciferum Boryangraecum A.cf.pinifolium A.pinifolium A.baronii A.ochraceum A.chaetopodium A.setipes A.caulescens A.cf.zeylanicum A.huntleyoides A.calceolus A.multiflorum A.cf.multiflorum A.moratii A.serpens A.florulentum A.penzigianum Robusta A.oblongifolium A.amplexicaule A.eburneum var.superbum A.eburneum var.longicalcar Angraecum A.eburneum var.xerophilum A.eburneum var.eburneum A.polystachyum Oeoniella A.conchoglossum A.corrugatum A.linearifolium A.conchiferum A.expansum A.mirabile A.arachnites Arachnangraecum A.germinyanum A.cf.germinyanum A.appendiculatum A.pseudofilicornu A.scottianum A.sedifolium A.triangulifolium Angraecoides A.pingue A.nanum Nana A.dives A.teres Africanae A.clareae A.compactum A.leonis Humblotiangraecum A.magdalenae A.viguieri A.bracteosum A.striatum Hadrangis A.cadetii A.humbertianum Sobennikoffia A.sesquipedale A.sororium Stellariangraecum A.ankeranense A.didieri A.elephantinum A.pseudodidieri A.cf.liliodorum A.liliodorum A.obesum A.cornigerum A.letouzeyi A.cf.rutenbergianum A.sterophyllum Perrierangraecum A.cf.elephantinum A.dryadum A.borbonicum A.cf.breve A.cf.peyrotii A.peyrotii A.cucullatum A.drouhardi A.lecomtei A.praestans A.danguyanum A.mauritianum A.pectinatum A.ramosum A.cf.panicifolium Pseudojumellea A.dasycarpum A.filicornu A.cf.humblotianum A.panicifolium Lemurorchis madagascariensis Cryptopus elatus Cryptopus paniculatus Oeonia rosea Acaulia rhynchoglossa Neobathiea grandidierana Lemurella pallidiflora Lemurella papillosa Parangraecum cf.humile Parangraecum pterophyllum Parangraecum perparvulum Erasanthe henricii Beclardia macrostachya

Fourteen clades (Fig. 2.1) are recognized here as sections of Angraecum, including nine sections of Garay (1973): Angraecoides Schltr., Angraecum Bory, Arachnangraecum Schltr., Boryangraecum Schltr., Hadrangis Schltr., Humblotiangraecum Schltr., Nana (Cordem.) Garay, Perrierangraecum Schltr. and Pseudojumellea Schltr. Five new sections are defined following our phylogenetic analysis (chap. 1): Africanae Andriananjamanantsoa, Oeoniella Andriananjamanantsoa, Robusta Andriananjamanantsoa, Sobennikoffia Andriananjamanantsoa and Stellariangraecum Andriananjamanantsoa. Modifications were made to published descriptions to better describe the sections. Sections Oeoniella and Sobennikoffia are new combinations for the genera Oeoniella and Sobennikoffia, fully embedded within Angraecum (Fig. 2.1), as confirmed by phylogenetic studies (Carlsward et al., 2006; chap. 1). Andriananjamanantsoa et al. (chap. 1) mentioned the morphological resemblance between these genera and Angraecum, despite the differences in the form of the labellum highlighted by Schlechter (1918, 1925). In this treatment, all species that have been sampled in the phylogenetic reconstructions are presented in bold.

2.3. Sections of Angraecum

2.3.1. Angraecum Bory, Voy. 1: 359, t. 19 (1804).

TYPE: Angraecum eburneum Bory

Epiphytic, lithophytic, rarely terrestrial plants; stem erect, commonly short, or elongate, rarely reduced; leaves distichous, articulate with leaf-sheaths commonly distinct, sometimes densely imbricate; inflorescence axillary, sessile or pedunculate, one-flowered or racemose, usually one per node; flowers sessile or pedicellate, alternate or secund along rachis, sepals and petals free, spreading, more or less reflexed; labellum lowermost or uppermost, sessile, entire, more or less concave, spreading, basally enveloping the column, rarely lobed in the apex; spur distinct, very short (less than 5 mm) to very long (up to 35 cm); column fleshy, very short, without a foot; clinandrium rather shallow, in front deeply bilobed with an abbreviated, tooth-like rostellum in the middle; pollinia 2, globose, sulcate, attached to a common viscidium or each pollinium may be attached more or less to a separate viscidium;

ovary round in section sometimes angular, resupinate from 180° to 360°, rarely non- resupinate.

2.3.2. Angraecum sect. Africanae Andriananjamanantsoa sect. nov.

TYPE: Angraecum teres Summerh.

DISTRIBUTION: Kenya, Socotra, Somalia, Tanzania

Epiphytic, medium or large plants (to 25 cm height); stems erect, short, internodes short; leaves densely imbricate, coriaceous; inflorescences racemose, many-flowered (up to 10), infrafoliar, longer or equaling leaves, one to four per stem; flowers small, greenish, sepals and petals linear-lanceolate; labellum uppermost, navicular, shorter or equaling dorsal sepal, apex attenuate; spur longer or equaling dorsal sepal, isodiametric; ovary round in section with 360° resupination.

ETYMOLOGY: The name refers to the geographical distribution of the species that are restricted to Eastern Africa (Stewart et al., 2006).

This section groups all African species placed previously in section Boryangraecum. Further phylogenetic work is required to confirm the placement of A. geniculatum G.Will. suggested here. This species from Zambia was accommodated in section Boryangraecum by Garay (1973). Indeed, species in section Africanae are very similar to those in section Boryangraecum Schltr. sensu Garay found in Madagascar and the western Indian Ocean Island. Geographical distribution seems to have an effect on genetic drift which is so far the most plausible explanation of the segregation of the species. The other explanation could be convergence evolution, which seems frequent in the Angraecinae (chap. 1).

Species included

Angraecum dives Rolfe, A. geniculatum G.Will., A. teres Summerh.

2.3.3. Angraecum sect. Angraecoides (Cordem.) Garay, Kew Bull. 28: 495– 516 (1973).

LECTOTYPE: Angraecum pingue Frapp., designated by Garay, Kew Bull. 28: 495–516 (1973).

Mystacidium sect. Angraecoides Cordem. Rev. Gin. Bot. II: 413 (1899).

Angraecoides (Cordem.) Szlach., Mytnik & Grochocka, Biodiv. Res. Conserv. 29: 1–23 (2013).

DISTRIBUTION: Madagascar, Mauritius, Reunion

Epiphytic on trunk, medium or large plants (to 25 cm height); stems erect, elongate, internodes short; leaves alternate spiral, often succulent; inflorescences always one-flowered, subterminal, longer than leaves, one to four per stem; flowers small, greenish or greenish yellow, sepals and petals linear-lanceolate; labellum lowermost, concave or rounded at the base, shorter or equaling dorsal sepal, rostrate-acuminate; spur longer or equaling dorsal sepal, with a wide opening, isodiametric; ovary round in section with 180° resupination.

We restrict section Angraecoides to species with a lowermost labellum. The African genus Angraecoides (Cordem.) Szlach., Mytnik & Grochocka (former Angraecum sect. Conchoglossum, endemic to Africa) needs a new name since the type species of this genus and of section Angraecoides, A. pingue, belongs to Angraecum sect. Angraecoides according to phylogenetic data (Fig. 2.1, chap. 1) and not to the segregate entity.

Species included

Angraecum pingue Frapp. ex Cordem., A. sedifolium Schltr., A. triangulifolium Senghas.

2.3.4. Angraecum sect. Angraecum Benth., Benth. & Hook. F. Gen. Pl. 3: 583 (1883).

TYPE: Angraecum eburneum Bory

DISTRIBUTION: Comoros, Kenya, Madagascar, Mascarenes, Seychelles, Tanzania

Epiphytic or lithophytic, very large plants (up to 1 m height); stems short to elongate, internodes long to 5 cm when existing; leaves densely imbricate, coriaceous; inflorescences racemose, suprafoliar, longer than leaves, many-flowered (up to 10), one to five per stem; flowers very large, sepals and petals linear-lanceolate, green; labellum uppermost, orbicular concave, white, apex acuminate; spur medium to very long (up to 45 cm), green, isodiametric from the base, tapering towards the tip; ovary round in section with 360° resupination.

This section is restricted to A. eburneum and its varieties. It is strongly supported in morphological and molecular phylogenetic analyses (chap. 1). Previous descriptions were based on the significantly large size of the plants and flowers (Schlechter, 1925; Perrier de la Bathie, 1941; Garay, 1973; Stewart et al., 2006; Cribb et Hermans, 2009) without taking in consideration the position of the labellum (uppermost vs lowermost) or the number of flower per inflorescence (racemose vs few-flowered inflorescences). The taxonomic disposition of A. longicalcar (Bosser) Senghas raises question about its recognition as a distinct species versus a variety of A. eburneum as proposed by Bosser (1965). Further molecular phylogenetic and morphometric studies are required before making a decision as to its exact position.

Species included

A. eburneum Bory var. eburneum, A. eburneum var. giryamae (Rendle) Senghas & P.J.Cribb, A. eburneum var. longicalcar Bosser, A. eburneum var. superbum (Thouars) H.Perrier, A. eburneum var. xerophilum H.Perrier.

2.3.5. Angraecum sect. Arachnangraecum Schltr., Fedde, Repert. Sp. Nov. Beih. 33: 309 (1925).

LECTOTYPE (in hoc loco electus): Angraecum arachnites Schltr.

DISTRIBUTION: Comoros, Kenya, Malawi, Madagascar, Mascarenes, Mozambique, South Africa, Tanzania, Zimbabwe

Epiphytic on trunk, large plants (to 30 cm height); stems elongate, pendent, often with long internode (up to 5 cm); leaves alternate, coriaceous; inflorescences one-flowered,

subterminal, longer than leaves, one to three per stem; flowers medium to large, sepals and petals tubular, arachniform, white green sometimes ocher; labellum uppermost, rather broad, more or less suborbicular in outline, concave, white, apex acuminate; spur elongate to 13 cm, sigmoid, greenish white, sometimes ocher, slightly large at the base, tapering towards the tip; ovary round in section with 360° resupination.

The lectotype designated by Garay (1973), A. ramosum Thouars, was misplaced in this section and has been moved to section Pseudojumellea (Micheneau et al., 2008a). The type specimen is in fruit which is in fact confusing because the habit and the inflorescence look very similar to sect. Arachnangraecum (stems elongate, inflorescence longer than leaves, pedicellate ovary round in section). This might be the reason why Perrier de la Bathie (1941) transferred several species of sect. Arachnangraecum as varieties of A. ramosum (e.g. A. arachnites, A. conchoglossum and A. germinyanum) and later why Garay (1973) chose A. ramosum as type of the section. The major differences lie on the flower shape, where the arachniform, uppermost labellum and the sigmoid spur are typical of section Arachnangraecum but are not present in A. ramosum. Therefore, we are designating a new lectotype to represent the section as originally described by Schlechter (1925), with a modification to the original description following recent morphological and phylogenetic evidence (chapter 1).

Species included

Angraecum ampullaceum Bosser, A. appendiculatum Frapp. ex Cordem., A. arachnites Schltr., A. conchiferum Lindl., A. conchoglossum Schltr., A. corrugatum (Cordem.) Micheneau, A. expansum Thouars, A. germinyanum Hook.f., A. linearifolium Garay, A. mirabile Schltr., A. pseudofilicornu H.Perrier, A. scottianum Rchb.f., A. teretifolium Ridl.

2.3.6. Angraecum sect. Boryangraecum Schltr., Fedde, Repert. Sp. Nov., Beih. 33: 308 (1925).

TYPE: Angraecum pumilio Schltr.

Angraecum sect. Acaulia Garay, Kew Bull. 28: 495–516 (1973). Type: Angraecum rhynchoglossum Schltr.

Angraecum sect. Chlorangraecum Schltr. Fedde, Repert. Sp. Nov., Beih. 33: 310 (1925). Lectotype: Angraecum chloranthum Schltr., designated by Garay, Kew Bull. 28: 495–516 (1973) [syn. of Angraecum huntleyoides Schltr., Bot. Jahrb. Syst. 38: 160 (1906)].

Angraecum sect. Gomphocentrum (Benth.) Garay, Kew Bull. 28: 495–516 (1973).

Mystacidium sect. Gomphocentrum Benth., J. Linn. Soc. Bot. 18: 337 (1881). Type: Angraecum caulescens Thouars

Angraecum sect. Lemurangis Garay, Kew Bull. 28: 495–516 (1973). Type: Angraecum madagascariense (Finet) Schltr. (= Macroplectrum madagascariense Finet).

Angraecum sect. Lepervenchea (Cordem.) Schltr., Beih. Bot. Centralbl. 36(2): 157 (1918).

Lepervenchea Cordem., Rev. Gin. Bot. II: 415 (1899). Type: Angraecum tenuifolium Frapp. ex Cordem.

DISTRIBUTION: Comoros, Madagascar, Mauritius, Mozambique, Reunion, Seychelles

Epiphytic on trunk, medium to large plants (up to 30 cm); stems very short to elongate, erect or pendent, internodes often long (to 5 cm) sometimes short or absent; leaves imbricate, thin; inflorescences one-flowered or racemose (up to 10 flowers), infra or suprafoliar, short sometimes longer than leaves, one to more than five per stem, sometimes branched; flowers small, yellowish green, sub-diaphanous, sepals and petals linear-lanceolate; labellum uppermost, navicular, shorter or longer than dorsal sepal, apex acute or acuminate; spur shorter or longer than dorsal sepal, often globular at the tip; ovary round in section with 360° resupination.

We modified the original description to better describe the section wich assembles here species previously placed in sections Acaulia, Angraecoides, Boryangraecum, Chlorangraecum, Gomphocentrum, Lemurangis and Lepervenchea by Garay (1973). Schlechter (1925) hesitated to separate sect. Hildebrandtiangraecum (sect. Gomphocentrum sensu Garay) and Micrangraecum (sect. Nana sensu Garay) from Boryangraecum because of their morphological resemblance (mostly floral). Phylogenetic reconstructions and

morphological data supported Nana as a separate section, as circumscribed by Garay (1973), which is phylogenetically distant from sect. Boryangraecum but closely related to sect. Africanae and Angraecoides (chapter 1). However, further molecular analyses, including a more complete sampling, will be required to confirm the composition of section Boryangraecum.

Species included

A. acutipetalum Schltr. var. acutipetalum, A. acutipetalum var. analabeensis H.Perrier ex Hermans, A. acutipetalum var. ankeranae H.Perrier ex Hermans, A. alleizettei Schltr., A. andasibeense H.Perrier, A. andringitranum Schltr., A. appendiculoides Schltr., A. aviceps Schltr., A. baronii (Finet) Schltr., A. brachyrhopalon Schltr., A. calceolus Thouars, A. caricifolium H.Perrier, A. caulescens Thouars, A. chaetopodum Schltr., A. chermezonii H.Perrier, A. cilaosianum (Cordem.) Schltr., A. cordemoyi Schltr., A. cornucopiae H.Perrier, A. corynoceras Schltr., A. costatum Frapp. ex Cordem., A. crassifolium Schltr., A. curvicaule Schltr., A. curvipes Schltr., A. dauphinense (Rolfe) Schltr., A. decaryanum H.Perrier, A. dupontii Pailler, A. falcifolium Bosser, A. ferkoanum Schltr., A. flavidum Bosser, A. floribundum Bosser, A. hermannii Schltr., A. huntleyoides Schltr., A. inapertum Thouars, A. madagascariense (Finet) Schltr., A. multiflorum Thouars, A. muscicolum H.Perrier, A. musculiferum H.Perrier, A. myrianthum Schltr., A. obversifolium Frapp. ex Cordem., A. ochraceum (Ridl.) Schltr., A. oeonioides Bosser, A. onivense H.Perrier, A. parvulum Ayres ex S.Moore, A. patens Frapp. ex Cordem., A. pauciramosum Schltr., A. pergracile Schltr., A. perhumile H.Perrier, A. pinifolium Bosser, A. pseudopetiolatum Frapp. ex Cordem., A. pumilio Schltr., A. rhizanthium H.Perrier, A. rhizomaniacum Schltr., A. rostratum Ridl., A. sacciferum Lindl., A. sacculatum Schltr., A. salazianum Schltr., A. scalariforme H.Perrier, A. setipes Schltr., A. sinuatiflorum H.Perrier, A. tamarindicolum Schltr., A. tenuifolium Frapp. ex Cordem., A. tenuipes Summerh., A. tenuispica Schltr., A. undulatum (Cordem.) Schltr., A. verecundum Schltr., A. vesiculatum Schltr., A. vesiculiferum Schltr., A. viridiflorum Cordem., A. xylopus Rchb.f., A. zaratananae Schltr., A. zeylanicum Lindl.

2.3.7. Angraecum sect. Hadrangis Schltr., Beih. Bot. Centralbl. 36(2): 158 (1918).

TYPE: Angraecum striatum Thouars

DISTRIBUTION: Mauritius, Reunion

Epiphytic on canopy or trunk, large plants (to 35 cm height); stems erect, short or elongate, internodes short; leaves dense, coriaceous; inflorescences racemose, few to several- flowered (up to 10), suprafoliar, short or equaling leaves, one to four per stem; flowers medium, white, sepals and petals lanceolate-ovate; labellum lowermost, ovate, deeply concave, shorter or equaling dorsal sepal, white overlaid green at the moth, apex acute; spur shorter than dorsal sepal, conic or saccate, with a wide opening, green; ovary round in section with 180° resupination.

This is the only section which both was recognized as a natural group from its original description (Schlechter, 1918) and supported in morphological and molecular phylogenetic studies (chap. 1).

Species included

Angraecum bracteosum Balf. & S.Moore, A. cadetii Bosser, A. jeannineanum Fournel & Micheneau, A. striatum Thouars.

2.3.8. Angraecum sect. Humblotiangraecum Schltr., Fedde, Repert. Sp. Nov., Beih. 33: 310 (1925).

TYPE: Angraecum leonis (Rchb. f.) André [Aeranthes leonis Rchb. f.]

DISTRIBUTION: Madagascar, Comoros

Epiphytic on canopy or trunk, medium to large plants (to 35 cm height); stems erect, short or barely developed, internodes short to elongated (to 5 cm); leaves thin to thick, coriaceous or fleshy; inflorescences racemose, with up to four flowers, suprafoliar, longer than leaves, one to four per stem; flowers very large, white, sepals and petals linear-lanceolate to

lanceolate-ovate; labellum lowermost, broadly ovate, deeply concave, longer or equaling dorsal sepal, sometimes white overlaid green at the moth, apex acute; spur long (to 15 cm), with a wide often infundibuliform opening, tapering from the median to tip, greenish white towards the tip; ovary triquetrous or round in section with 180° resupination.

Some species previously placed in section Perrierangraecum (e.g. A. clareae and A. compactum) fit better in this section. Their position is supported by phylogenetic evidence (chap. 1). The triquetrous ovary, most often observed in sect. Perrierangraecum, and the plant habit, similar to some species from this section, might be reasons why these species were misplaced. The spur shape of A. praestans is remarkably similar to that of A. viguieri and this is probably one reason why Garay (1973) included this species in section Perrierangraecum.

Species included

Angraecum aloifolium Hermans & P.J.Cribb, A. clareae Hermans, la Croix & P.J.Cribb, A. compactum Schltr., A. dollii Senghas, A. leonis (Rchb.f.) André, A. magdalenae Schltr. & H.Perrier var. magdalenae, A. magdalenae var. latiilabellum Bosser, A. viguieri Schltr.

2.3.9. Angraecum sect. Nana (Cordem.) Garay, Kew Bull. 28: 495–516 (1973).

Mystacidium sect. Nana Cordem., Rev. Gin. Bot. II: 414 (1899).

LECTOTYPE: Angraecum nanum Frapp., designated by Garay, Kew Bull. 28: 495–516 (1973).

Angraecum sect. Micrangraecum Schltr., Fedde, Repert. Sp. Nov., Beih. 33: 308 (1925). Type: Angraecum pusillum Lindl.

DISTRIBUTION: South Africa, Ethiopia, Madagascar, Mauritius, Kenya, Reunion, Tanzania, Zambia, Zimbabwe

Epiphytic, tiny plants (to 5 cm heigth); stems short, internodes reduced; leaves thin, graminiform; inflorescences often one-sided racemose, infrafoliar, longer than leaves, one to three per stem; flowers minute, greenish, sepals and petals lanceolate-ovate; labellum

uppermost, navicular, apex acute; spur shorter than dorsal sepal, isodiametric; ovary round in section with 360° resupination.

We modified the original description and restricted the section to species with racemose inflorescences of minute, greenish flowers. Garay (1973) included all dwarf plants that bear minute flowers in this section, even though their vegetative and floral traits differed. All dwarf plants with thick leaves and small flowers (A. cf. humile, A. perparvulum, A. pterophyllum) are now placed in genus Parangraecum. Further phylogenetic and morphological studies, including a wider sampling of species, are required to validate the section because we only sampled A. nanum in the analyses. We suspect that some of the species in this section could be varieties or synonyms because the morphological differences are few (tiny plants, grass-like, inflorescence and flowers very similar).

Species included

Angraecum chamaeanthus Schltr., A. decipiens Summerh., A. microcharis Schltr., A. minus Summerh., A. minutum Frapp. ex Cordem., A. nanum Frapp. ex Cordem., A. pusillum Lindl.

2.3.10. Angraecum sect. Oeoniella (Schltr.) Andriananjamanantsoa comb. nov.

Oeoniella Schltr. Beih. Bot. Centralbl. 36(2): 176 (1918).

TYPE: Angraecum polystachyum (Thouars) A. Rich. [Oeoniella polystachys (Thouars) Schltr.]

DISTRIBUTION: Comoros, Madagascar, Mascarenes, Seychelles

Epiphytic on trunk or lithophytic undergrowth, medium plants (to 15 cm height); stems elongate, erect or climbing, internodes long (to 5 cm); leaves subimbricate, coriaceous; inflorescences racemose, suprafoliar, longer than leaves, many-flowered (up to 10), one to four per stem; flowers medium sized, sepals and petals linear-lanceolate, green; labellum uppermost, orbicular, rotund at the base, white, apex acuminate; spurs very short, green, isodiametric; ovary round in section, non-resupinate.

We believe that Oeoniella aphrodite could be a variety of Oeoniella polystachyus as they merely differ in size (Baker, 1877), which could be interpreted also as phenotypic variation: O. aphrodite appears larger than O. polystachys. Further investigations involving molecular and morphological work (including morphometry) are required before making any decision.

Species included and transfered from Oeoniella to Angraecum

Angraecum aphrodite (Balf. & S. Moore) Andriananjamanantsoa comb. nov.

Basionym: Listrostachys aphrodite Balf. & S. Moore, J.G. Baker, Fl. Mauritius: 354 (1877).

Synonym: Oeoniella aphrodite (Balf.f. & S.Moore) Schltr., Beih. Bot. Centralbl. 36(2): 177 (1918).

Angraecum polystachyum (Thouars) A. Rich., Mém. Mus. Hist. Nat. 4: 66 (1818).

Basionym: Epidendrum polystachys Thouars, Hist. Orchid.: t. 82 (1822).

Synonym: Oeoniella polystachys (Thouars) Schltr., Beih. Bot. Centralbl. 36(2): 177 (1918).

2.3.11. Angraecum sect. Perrierangraecum Schltr., Fedde, Repert. Sp. Nov., Beih. 33: 309 (1925).

LECTOTYPE (in hoc loco electus): Angraecum rutenbergianum Kraenzl.

DISTRIBUTION: South Africa, Comoros, Madagascar, Malawi, Reunion, Zimbabwe

Epiphytic on canopy or trunk, medium to large plants (to 35 cm height); stems erect, short to elongate, internodes very short rarely long (to 5 cm); leaves fleshy, usually thick, fan shape; inflorescence usually one-flower, sometimes racemose with up to four flowers, suprafoliar, very short, one or two per stem; flowers medium to large, white, sepals and petals linear-lanceolate to lanceolate-ovate; labellum lowermost, broadly ovate, equal to dorsal sepal, slightly concave, apex acute; spur short to elongate (to 15 cm), isodiametric at base, tapering to tip, greenish white towards the tip; ovary round in section often triquetrous with 180° resupination.

We designated a new type for the section since the one proposed by Garay (1973), Angraecum triquetrum, is a synonym of Jumellea triquetra (Thouars) Schltr. In the original description, Schlechter (1925) kept this species in section Perrierangraecum even though he already transfered it to Jumellea. The presence of Angraecum praestans in this section is difficult to accept without phylogenetic support (chap. 1) because the similarities of this species to species of sect. Humblotiangraecum are intriguing (the labellum and spur shapes in A. praestans and A. viguieri appear very similar). Indeed, the infudibular form of the spur is unusual in section Perrierangraecum. This form appears to represent an example of convergent evolution within Angraecum. We also added A. cornigerum (section Filangis) to this section because its morphological characters fit well here, a fact that is strongly supported by the phylogeny.

Species included

Angraecum ambrense H.Perrier, A. ankeranense H.Perrier, A. bicallosum H.Perrier, A. borbonicum Bosser, A. breve Schltr., A. chimanimaniense G.Will., A. clavigerum Ridl., A. compressicaule H.Perrier, A. coriaceum (Thunb. ex Sw.) Schltr., A. cornigerum Cordem., A. coutrixii Bosser, A. crassum Thouars, A. cucullatum Thouars, A. curnowianum (Rchb.f.) T.Durand & Schinz, A. curvicalcar Schltr., A. didieri (Baill. ex Finet) Schltr., A. drouhardii H.Perrier, A. dryadum Schltr., A. elephantinum Schltr., A. elliotii Rolfe, A. equitans Schltr., A. humbertii H.Perrier, A. imerinense Schltr., A. kranzlinianum H.Perrier, A. lecomtei H.Perrier, A. letouzeyi Bosser, A. liliodorum Frapp. ex Cordem., A. litorale Schltr., A. longicaule Humbert, A. mahavavense H.Perrier, A. meirax (Rchb.f.) H.Perrier, A. obesum H.Perrier, A. palmicolum Bosser, A. peyrotii Bosser, A. praestans Schltr., A. pseudodidieri H.Perrier, A. rigidifolium H.Perrier, A. rutenbergianum Kraenzl., A. sambiranoense Schltr., A. stella-africae P.J.Cribb, A. sterrophyllum Schltr., A. trichoplectron (Rchb.f.) Schltr., A. urschianum Toill.-Gen. & Bosser.

2.3.12. Angraecum sect. Pseudojumellea Schltr., Beih. Bot. Centralbl. 36(2): 157 (1918).

Angraecum sect. Filangis Garay, Kew Bull. 28: 495–516 (1973). Type: Angraecum filicornu Thou.

Angraecum sect. Pectinaria (Benth.) Schltr., Beih. Bot. Centralbl. 36(2): 157 (1918). Type: Angraecum pectinatum Thou.

Mystacidium sect. Pectinaria Benth., Benth. & Hook. f., Gen. P1. 3: 585 (1883).

LECTOTYPE: Angraecum mauritianum (Poir.) Frapp. [Orchis mauritiana Poir.], designated by Garay, Kew Bull. 28: 495–516 (1973).

DISTRIBUTION: Madagascar, Mauritius, Reunion

Epiphytic on trunk or undergrowth lithophytic, medium to large plants (to 1 m height); stems erect, climbing or pendent, internodes usually long (to 5 cm); leaves thin or coriaceous, stems and leaves usually spotted black; inflorescences one-flower, subterminal, short sometimes longer than leaves, one to five per stem; flowers small to medium sized, white, sepals and petals linear-lanceolate to elliptic; labellum lowermost, linear-lanceolate to elliptic, concave, longer or equaling dorsal sepal, apex acute; spur short to elongate (to 12 cm), isodiametric at base, tapering to tip, white sometimes greenish white towards the tip; ovary round in section with 180° resupination.

We modified the original description to better match the phylogenetic composition of the section. Two sections of Garay (1973) are subsumed within it: Pectinaria and Pseudojumellea. As originally proposed by Schlechter (1918), we moved A. filicornu (the type of section Filangis) to section Pseudojumellea, since this position is strongly supported by the phylogeny (chap. 1).

Species included

Angraecum danguyanum H.Perrier, A. darainense P.J. Cribb & Nusb., A. dasycarpum Schltr., A. filicornu Thouars, A. humblotianum Schltr., A. implicatum Thouars, A.

mauritianum (Poir.) Frapp., A. melanostictum Schltr., A. panicifolium H.Perrier, A. pectinatum Thouars, A. platycornu Hermans, P.J. Cribb & Bosser, A. ramosum Thouars.

2.3.13. Angraecum sect. Robusta Andriananjamanantsoa sect. nov.

TYPE: Angraecum oblongifolium Toill.-Gen. & Bosser

DISTRIBUTION: Madagascar, Comoros

Epiphytic on trunk, medium to very large plants (up to 1 m height); stems elongate, pendent, robust, often with long internode (up to 5 cm); leaves alternate, often coriaceous; inflorescences one to three-flowered, subterminal, shorter than leaves, one to five per stem; flowers medium, white, sepals and petals lanceolate-ovate; labellum uppermost, usually smaller than dorsal sepal, elliptic to lanceolate, concave, apex acuminate; spur short to elongate (to 12 cm), sigmoid, slightly enlarged at the base, tapering towards the tip, green; ovary round in section with 360° resupination.

ETYMOLOGY: The name refers to the structure of the stem, which is very robust and woody- like.

This section looks very similar to section Arachnangraecum, except that the flowers are smaller and the margin of sepals and sepals always entire. The species in this section were previously placed in section Angraecum (A. dendrobiopsis, A. penzigianum, A. serpens) because of their racemose inflorescence, and Pseudojumellea (A. oblongifolium) because of their medium-sized flowers and one-flowered inflorescence. We also added A. amplexicaule and A. moratii previously placed in section Filangis to this section because of morphological characters and phylogenetic evidence (Fig. 2.1).

Species included

Angraecum amplexicaule Toill.-Gen. & Bosser, A. dendrobiopsis Schltr., A. florulentum Rchb.f., A. moratii Bosser, A. oblongifolium Toill.-Gen. & Bosser, A. penzigianum Schltr., A. serpens (H.Perrier) Bosser.

2.3.14. Angraecum sect. Sobennikoffia (Schltr.) Andriananjamanantsoa comb. nov.

Sobennikoffia Schltr., Repert. Spec. Nov. Regni Veg. Beih. 33: 361 (1925).

TYPE: Sobennikoffia robusta (Schltr.) Schltr. [Oeonia robusta Schltr.]

DISTRIBUTION: Western Madagascar

Epiphytic on canopy or trunk, large plants (to 40 cm height); stems erect, short or elongate, internodes short; leaves dense, coriaceous; inflorescences racemose, many-flowered (up to 10), suprafoliar, longer than leaves, one to four per stem; flowers large, white, sepals and petals lanceolate-ovate; labellum lowermost, ovate, deeply concave, longer or equaling dorsal sepal, white overlaid green at the moth, apex trilobed; spur longer or equaling dorsal sepal, infundibuliform with a wide opening, green; ovary round in section with 180° resupination.

In the phylogeny (Fig. 2.1; chap. 1), sect. Sobennikoffia is sister to sections Hadrangis and Humblotiangraecum. Morphologically, they share several traits: large flowers, lowermost labellum, broadly ovate and deeply concave, often white overlaid green at the moth, and spurs widely open at the base. This section is widespread from north western to south western Madagascar.

Species included and transferred from Sobennikoffia to Angraecum

Angraecum fournierianum Kraenzl, Gard. Chron. 3(15): 808 (1894).

Synonym: Sobennikoffia fournieriana (Kraenzl.) Schltr., Repert. Spec. Nov. Regni Veg. Beih. 33: 362 (1925).

Angraecum humbertianum (H.Perrier) Andriananjamanantsoa comb. nov.

Basionym: Sobennikoffia humbertiana H.Perrier, Notul. Syst. Paris, 7: 134 (1938).

Angraecum poissonianum (H.Perrier) Andriananjamanantsoa comb. nov.

Basionym: Sobennikoffia poissoniana H.Perrier, Notul. Syst. Paris, 14: 164 (1951).

Angraecum robustum (Schltr.) Schltr., Beih. Bot. Centralbl. 33(2): 437 (1915).

Basionym: Oeonia robusta Schltr., Ann. Inst. Bot.-Géol. Colon. Marseille, 3(1): 184, t. 13 (1913).

Synonym: Sobennikoffia robusta (Schltr.) Schltr., Repert. Spec. Nov. Regni Veg. Beih. 33: 362.

2.3.15. Angraecum sect. Stellariangraecum Andriananjamanantsoa sect. nov.

TYPE: Angraecum sororium Schltr.

DISTRIBUTION: Madagascar

Epiphytic or lithophytic, very large plants (to 50 cm height); stems erect, internodes very short; leaves densely imbricate, coriaceous; inflorescences usually one or two-flowered, sometimes racemose with up to five flowers, suprafoliar, short or equaling leaves, one to four per stem; flowers very large, white or cream, sepals and petals linear-lanceolate to lanceolate- ovate; labellum lowermost, broadly ovate, slightly concave, longer or equaling dorsal sepal, apex acute; spur very long (up to 35 cm), isodiametric from the base, tapering towards the tip, pale green; ovary round in section with 180° resupination.

ETYMOLOGY: The name comes from a local appellation, ‘Comet Orchid’, which designates Angraecum species (A. sesquipedale and A. sororium) that have large, white flowers with a long spur.

Two species previously placed in sect. Angraecum, A. sesquipedale and A. sororium, are accommodated in this section according to both morphology and molecular evidence (Fig. 2.1). The lowermost position of the labellum and the untwisted ovary are the major characters that separate species in this section from section Angraecum (its previous section). A third species, Angraecum protensum Schltr., which was not included in previous phylogenetic studies, is referable to this section. This species has a very similar habit to A. sororium and inhabits limestone inselbergs in the Middle West central highland of Madagascar.

Species included

Angraecum protensum Schltr., A. sesquipedale Thouars var. sesquipedale, A. sesquipedale var. angustifolium Bosser & Morat, A. sesquisectangulum Kraenzl., A. sororium Schltr.

2.3.16. Incertae sidis

Two species poorly know from herbaria specimen should be removed from Angraecum:

Angraecum gracile Thouars, Hist. Orchid.: t. 77 (1822). Garay (1973) and Stewart et al. (2006) consider this endemic species to Mauritius a synonymy of Chamaeangis gracilis (Thouars) Schltr., Beih. Bot. Centralbl. 36(2): 108 (1918).

Angraecum palmiforme Thouars, Hist. Orchid.: 68. (1822). Garay (1973) putted this species in section Angraecum, while Stewart et al. (2006) believe this endemic species to Reunion is probably extinct.

2.3.17. Species excluded from Angraecum

The following species were misidentified and are relegated to synonymy:

Angraecum metallicum Sander = Aerangis stylosa (Rolfe) Schltr., Beih. Bot. Centralbl. 33(2): 427 (1915).

Angraecum ramulicolum H.Perrier = Aerangis pallidiflora H.Perrier, Notul. Syst. (Paris) 7: 36 (1938).

Angraecum triquetrum Thouars = Jumellea triquetra (Thouars) Schltr., Beih. Bot. Centralbl. 33(2): 430 (1915).

Taking into account all previous phylogenetic studies (Carlsward et al., 2006; Micheneau et al., 2008a; Szlachetko et al., 2013; chap. 1), all African sections are excluded from Angraecum. Nonetheless, further studies with a more complete sampling are needed before attributing unplaced species to genus, especially within the former section Afrangraecum sensu Garay (1973) which includes ten species restricted to continental Africa.

In fact, only one species of this section (A. multinominatum) has been sampled in a phylogenetic reconstruction (Szlachetko et al., 2013) that appeared to be related to African Angraecinae genera rather than Angraecum sensu stricto, but it has not been attributed to any genera yet.

2.4. New genera of Malagasy orchids

2.4.1. Acaulia (Garay) Andriananjamanantsoa comb. nov.

TYPE: Acaulia rhynchoglossa (Schltr.) Andriananjamanantsoa [=Angraecum rhynchoglossum Schltr.]

Angraecum sect. Acaulia Garay, Kew Bull. 28: 495–516 (1973). Type: Angraecum rhynchoglossum Schltr.

DISTRIBUTION: Central highland and eastern Madagascar

Epiphytic plants, small, stemless; leaves 2-6, distinctly lanceolate-ovate, constricted at the base, 23 x 5 mm; inflorescences 16 mm long, 1-flowered (a second aborted when present), equaling or exceeding leaves; flowers small, pale green, petals and sepals linear-lanceolate 0.6 x 7.5 mm, dorsal sepal slightly longer; labellum uppermost 6 x 3 mm, base expanded and very rounded, apex acuminate; spur funnel-shaped and tapering below, often expanding a little at the apex, at least twice longer than labellum, ca. 25 mm long; ovary round in section with 360° resupination.

HABITAT: At low altitudes, usually on small trees or twigs (1 to 2 m from the ground), frequent in secondary forests.

CONSERVATION STATUS: Data deficient but possibly threatened because of its limited range.

Plants of this monotypic genus have a habit very similar to Lemurella Schltr., but the leaves are narrower, the stem reduced and the inflorescence one-flowered (Fig. 2.2). It differs from other species that were included in Angraecum section Acaulia by Garay (1973) in its

very acuminate labellum and linear sepals and petals. These other species are now part of section Boryangraecum.

Fig. 2.2. Acaulia rhynchoglossa (Photo: Andriananjamanantsoa)

SPECIMEN EXAMINED: Madagascar, Antananarivo Prov., Vohitrilongo mountain, Perrier 14973 (holotype, https://science.mnhn.fr/institution/mnhn/collection/p/item/p00098421)

New species combination

Acaulia rhynchoglossa (Schltr.) Andriananjamanantsoa comb. nov.

Angraecum rhynchoglossum Schltr., Repert. Spec. Nov. Regni Veg. Beih. 33: 339 (1925).

2.4.2. Parangraecum Andriananjamanantsoa gen. nov.

TYPE: Parangraecum pterophyllum (H.Perrier) Andriananjamanantsoa [= Angraecum pterophyllum H.Perrier]

DISTRIBUTION: Eastern Africa (Ethiopia, Kenya, Rwanda, Swaziland, Tanzania, Zimbabwe), Madagascar (Central highland and Eastern), Reunion

Leaves elliptic or linear, semi-terete or slightly triquetrous, inflorescences racemose, flowers almost actinomorphic, labellum similar to petals but with raised edges, gynostemium pale green, spur globular.

Epiphytic plants, small and erect; stems usually elongate (up to 5 cm long), but very short or even inexistent in some species; leaves 2-10, inflorescences 1--5-flowered, flowers mostly one-sided, borne from lower leaf axils; flowers white; sepals ovate (3 x 2 mm), apex attenuate; petals similar to sepals, smaller (2.5 x 1.8 mm); labellum uppermost, ovate-concave (2.5 x 1.8 mm), white overlain with green at base; spur shorter than sepal (2 x 1.2 mm); ovary glabrous, 2 mm long.

HABITAT: Humid and evergreen forests, on trunk or lower branches.

CONSERVATION STATUS: Not evaluated but probably of little concern because it is abundant in the field (personal observation).

ETYMOLOGY: For the morphology (vegetative and floral) that is similar to that of Angraecum.

The leaves are fleshy and the apex acute. The inflorescence is usually short and the flowers dense (Fig. 2.3). Three Angraecum species (A. humile, A. perparvulum, and A. pterophyllum), which form a monophyletic group, are included in this genus. Two other species (A. rubellum and A. tenellum), which were not included in phylogenetic studies, appear better accommodated in this genus; this needs to be validated with molecular data.

Fig. 2.3. Parangraecum cf. humile (Photo: Andriananjamanantsoa)

SPECIMEN EXAMINED: Madagascar, Toamasina Prov., Mantadia, Onive basin, Perrier 18647 (holotype, https://science.mnhn.fr/institution/mnhn/collection/p/item/p00334625).

New species combination

Parangraecum humile (Summerh.) Andriananjamanantsoa comb. nov.

Basionym: Angraecum humile Summerh., Kew Bull. 13: 269 (1958).

Parangraecum perparvulum (H.Perrier) Andriananjamanantsoa comb. nov.

Basionym: Angraecum perparvulum H.Perrier, Notul. Syst. (Paris) 7: 123 (1938).

Parangraecum pterophyllum H.Perrier Andriananjamanantsoa comb. nov.

Basionym: Angraecum pterophyllum H.Perrier, Notul. Syst. (Paris) 7: 106 (1938).

Parangraecum tenellum (Ridl.) Andriananjamanantsoa comb. nov.

Basionym: Mystacidium tenellum Ridl., J. Linn. Soc., Bot.21: 489 (1885).

Angraecum tenellum (Ridl.) Schltr., Beih. Bot. Centralbl. 33(2): 438 (1915).

Angraecum bemarivoense Schltr., Repert. Spec. Nov. Regni Veg. Beih. 33: 332 (1925).

Parangraecum rubellum (Bosser) Andriananjamanantsoa comb. nov.

Basionym: Angraecum rubellum Bosser, Adansonia, n.s. 10: 22 (1988).

2.5. Acknowledgements

The authors thank A. Schuiteman (Royal Botanic Gardens, Kew) and M. Pignal (Museum of Paris) for their help in accessing herbarium specimens. We are thankful to J. Saarela, S. Joly, K. Gandhi and J. Hermans for their comments, assistance and expertise. We would like to acknowledge Les Amis du Jardin Botanique de Montréal, the Institut de recherche en biologie végétale (IRBV), and the Quebec Center for Biodiversity Science (QCBS) for funds for fieldwork and support for conferences. This study was supported by the Omaha’s Henry Doorly Zoo and the Madagascar Biodiversity Partnership.

Chapitre 3: Biogeography of Angraecum (Orchidaceae, Angraecinae), out of Madagascar case-dispersal

Herinandrianina N. Andriananjamanantsoa and Luc Brouillet

Département de sciences biologiques, Université de Montréal, 90 Av. Vincent d’Indy, Montréal (QC) H2V 2S9, Canada

Journal ciblé: Journal of Biogeography (Modifié pour les besoins de la présente thèse)

Tous les auteurs ont donné leur autorisation pour ce manuscrit à être inclus dans cette thèse

Contribution respective des auteurs:

Tahiana Andriananjamanantsoa: Conception du projet, collecte, traitement et analyse des données, interprétation et rédaction

Luc Brouillet: Supervision, conception du projet, conseils, recommandations, corrections et direction générale

3.1. Résumé/Abstract

Résumé

Objectifs: Cette étude vise à étudier la répartition des espèces d'Angraecum dans le monde, un groupe d’orchidées très diversifié à Madagascar et dans la région occidentale de l’Océan Indien, et une partie de l'Afrique tropicale. Nous avons évalué l'origine biogéographique de la sous-tribu des Angraecinae, et évalué l'effet de changement de stratification sur la diversification au sein du genre Angraecum à Madagascar. Nous voulons aussi analyser la diversification des espèces d'Angraecum après un changement de la strate occupée dans l'habitat.

Localisation: Madagascar, région occidentale de l’Océan Indien, Afrique

Méthodes: Nous avons évalué différents modèles de transition de la stratification écologique qui pourrait affecter la diversification d'Angraecum en utilisant les modèles multistate speciation and extinction. Ensuite, à partir de la phylogénie moléculaire datée des Angraecinae (chap. 1), nous avons reconstruit les aires ancestrales de la sous-tribu des Angraecinae en utilisant le programme BioGeoBEARS.

Résultats: Le meilleur modèle de transition suggère que le passage des épiphytes de canopée vers des lithophytes sous canopée ou sur inselberg, ou vers des épiphytes de tronc est le changement d’habitat (strate) le plus probable ayant pu conduire à la diversification d'Angraecum. L'analyse biogéographique montre que Madagascar est le lieu d'origine de la sous-tribu des Angraecinae et du genre Angraecum à partir d'un ancêtre du sud-est asiatique. Les lignées ancestrales d'Angraecum proviendraient du Nord-Est de Madagascar et sont apparues au Miocène. Chez Angraecum, la migration à l'intérieur de Madagascar a commencé vers la fin du Miocène, alors que la migration hors de Madagascar a commencé plus tard, généralement au Pléistocène.

Conclusions: Notre étude a montré un nouveau cas de dispersion à partir de Madagascar pour la sous-tribu des Angraecinae et le genre Angraecum. Chez Angraecum, les évenements de dispersion ont eu lieu au cours du Pliocène-Pléistocène et ont débuté à partir du Nord-Est de Madagascar. Les cyclones jouent un rôle important dans la dispersion à longue distance des

orchidées dans la région occidentale de l'Océan Indien. La diversification d'Angraecum dans les îles Comores et les Mascareignes reflète l'importance des nouveaux habitats et la similarité des niches écologiques dans l'expansion des espèces.

Mots-clés: Angraecinae, Angraecum, histoire biogéographique, diversification, niche, nord est de Madagascar, Orchidaceae, taux de transition

Abstract

Aim: This study investigates the worldwide distribution of Angraecum, a highly diversified orchid group restricted to Madagascar, the western Indian Ocean, and parts of tropical Africa. We assessed the biogeographical origins of the Angraecinae and of Angraecum. Our goals were also to analyze the diversification of Angraecum species following changes in the strata occupied within the forest.

Location: Africa, Madagascar, western Indian Ocean region

Methods: Using a dated molecular phylogeny of subtribe Angraecinae (chapter 1), we performed an ancestral area reconstruction of subtribe Angraecinae as implemented in BioGeoBEARS. We also tested different transition models related to forest habitat stratification, which may have affected the diversification of Angraecum, using the multistate speciation and extinction models.

Results: The best model suggested that transitions from canopy epiphytes to lithophytes (both undergrowth and inselberg), and to trunk epiphytes were the most probable shifts in stratification that might have led to the diversification in Angraecum. Historical biogeographic analysis showed that Madagascar was the place of origin of Angraecinae as well as Angraecum. Ancestral lineages of Angraecum came probably from the northern tropical lowland forest of Madagascar and arose during the Miocene. Within the genus, expansion within Madagascar started in the late Miocene, while dispersal out of Madagascar occurred mostly during the Pleistocene.

Main conclusion: Our study highlighted a new case of dispersal out of Madagascar in Angraecinae and Angraecum. In Angraecum, dispersal events took place during the Pliocene-

Pleistocene, and they started from the northeast Madagascar. Cyclones appeared to have played an important role in orchid seed dispersal in the western Indian Ocean region. The diversification of Angraecum in the Comoros and Mascarenes Islands reflects the importance of new habitats and niche similarities in range expansion.

Keywords: Angraecinae, Angraecum, historical biogeography, diversification, niche, northeast Madagascar, Orchidaceae, transition rates

3.2. Introduction

The Western Indian Ocean Region (WIOR) offers valuable insights for historical biogeography studies because of its biodiversity richness (Goodman and Benstead, 2005; Buerki et al., 2013), and due on one hand to islands dating from the Cenozoic (Madagascar and Seychelles), which may have experienced vicariance, and on the other to recently formed islands (Comoros and Mascarene), which could have been subject to dispersal events. Island biogeographical theory predicts that the largest proportion of lineages within an island region comes from the nearest island and/or mainland source (MacArthur and Wilson, 1967; Warren et al., 2010). Based on geological history, only dispersal can explain the presence of many oceanic island biotas (de Queiroz, 2005). Thus, Madagascar has been a major source of colonizing lineages for surrounding islands (Raxworthy et al., 2002; Weyeneth et al., 2008; Krüger et al., 2012; Buerki et al., 2013), especially for Comoros and the Mascarenes (Buerki et al., 2013). Many studies have revealed that Malagasy biotas originated from Africa (e.g. Yoder et al., 1996; Yoder et al., 2006; Bacon et al., 2015). However, the majority of lineages in Madagascar have post-Gondwanan origins (Vences, 2004; Yoder and Nowak, 2006), suggesting that dispersal played a more important role in their origins than vicariance (Warren et al., 2010). At the same time, it is now widely accepted that present-day levels of diversity and endemism in Madagascar may be due to a combination of divergence by vicariance and in situ diversification by post Cenozoic dispersers (Vences et al., 2001; Yoder and Nowak, 2006; Warren et al., 2010; Kuntner and Agnarsson, 2011; Strijk et al. 2012). Recent studies supported an Asian origin and affinities to many lineages within Madagascar (e.g. Battistini

and Richard-Vindard, 1972; Stoddart, 1984; Cheke and Hume, 2008; Buerki et al., 2013), opening up a new perspective on historical biogeography of the island and the WIOR.

Rapid radiation is often highlighted as a primary cause of the diversification Madagascar biotas (Janssen et al., 2008; Townsend et al., 2009; Vences et al., 2009; Anthony et al., 2010; Jonsson et al., 2012; Christidis et al., 2014), as well as the Mascarene lineages (Thébaud et al., 2009). This mechanism is usually considered a result of ecological opportunity and/or founder effect (Thébaud et al., 2009; Matzke, 2013a; Alsos et al., 2015). The main way to validate wether species diversity is caused by adaptive radiation or multiple dispersal events is through the use of species rich taxa in biogeographical analyses. Genus Angraecum is of particular interest because of its high diversification (chap. 1) and of its widespread geographical distribution in the Western Indian Ocean Region ranging from Africa to Sri Lanka. Diversification studies suggested that the great number of species within Angraecum resulted from a steady accumulation of lineages through-time (chap. 1) and not from a rapid radiation as suggested for many Malagasy taxa (e.g. Townsend et al., 2009; Vences et al., 2009). The significant number of species present in Madagascar and the Indian Ocean Islands (IOI: Comoros, Mascarenes, and Seychelles), especially the Comoros and Mascarenes, raised the hypothesis of a possible origin of Angraecum in Madagascar. The relatively young age of subtribe Angraecinae (Ramirez et al., 2007; Gustafsson et al., 2010, Iles et al., 2015), as well as Angraecum (Jacquemyn et al., 2005; chap. 1) eliminates a priori a post-Gondwanan vicariance hypothesis. Using Mascarene Angraecoid data, Micheneau et al. (2008a) suggested a Malagasy origin for Angraecinae. This study was conducted with a limited sample of taxa (mostly taxa from the Mascarenes). The absence of resolved generic and species relationships limits the ability to evaluate historical and geographical hypotheses. With a fully reconstructed phylogeny of Angraecum (chap. 1), it is now possible to evaluate the historical biogeographic pattern of the group.

The elevational gradient is often pointed out as the main cause of diversification in orchids (Jacquemyn et al., 2005; Acharya et al., 2011) and epiphytism was a key innovation (apomorphic state in Orchidaceae) responsible for a species burst in the family. Angraecum, with over 95% of its species epiphytic, took advantage of this innovation, being the most diverse genus in subtribe Angraecinae. The genus has mainly diversified in Madagascar and

the adjacent islands (Comoros and Mascarenes). Species diversification is often associated with historical biogeography and ecological niche opportunities (Schluter, 2000; Whittaker et al., 2008; Strijk et al., 2013). In fact, the origin of diversification of Angraecum in Madagascar was reported to be associated with paleoclimate events, essentially the establishment of the monsoon in the northern part of the island during the Miocene (Buerki et al., 2013), and the contraction and expansion of the tropical rainforest during the post-Pleistocene global cooling (Ruddiman, 2001). In Angraecum, we noted four types of habitat distribution according to the stratum (forest layer): in the canopy, on trunks, on undergrowth rocks, and on inselbergs (granite or limestone). Considering each layer as an ecological niche, the shift of stratification could be viewed as the colonization of a new niche. The ability of orchid seeds to occupy a habitat is determined by many factors such as niche opportunity, mycorrhizal source availability, or climatic conditions (Dressler, 1981–1993; Arditti and Ghami, 2000; Jersakkova and Malinova, 2007). Furthermore, adapted pollinators are required to maintain the population’s reproduction whatever the niche occupied. Understanding the mechanisms that help to colonize different strata within the forest will be helpful in understanding the distribution patterns of Angraecum in tropical rain forests.

Although the biogeography of subtribe Angraecinae has been studied previously (Micheneau et al., 2008), no study has included temporal events which are important to estimate the underlying causes of the distribution. Thus, many questions remain unanswered such as how, when and from where did the migration occur. Furthermore, the high number of Angreacum species encountered in Madagascar and surrounding islands (Mascarenes and Comoros) raises questions about the importance of ecological niches and its possible limitation on species distribution. We suspected that stratification shifts in the forest might play a significant role in colonizing new habitats. Using a dated and more thoroughly sampled phylogeny of Angraecinae (chap. 1), we are inferring the historical biogeography of Angraecum and its close relatives. Our goals were to: (1) evaluate the effect of forest stratification changes on the diversification of Angraecum, (2) evaluate the biogeographical origin of Angraecinae and its probable source, (3) assess biogeographic patterns within Angraecum, and (4) understand the mechanism of dispersal responsible for the distribution of the lineages.

3.3. Material and methods

3.3.1. Phylogenetic inference and divergence time estimates

Phylogenetic inference and divergence time estimates used in this analysis, as well as subsequent taxonomic changes, are described more fully in chap. 1 and 2. A total of 194 taxa were included in phylogenetic analyses, comprising 96 Angraecum species, 93 Angraecinae, and five outgroups (Acampe ochracea, Aerides odorata, Phalaenopsis cornu-cervi, Polystachya fulvilabia, and Vanda tricolor). Three plastid regions were combined (matK coding gene, rps16 intergenic spacer, and trnL intron) for a total of 3725 characters. DNA sequences were aligned using SeaView v4 (Gouy et al., 2010), and alignments were visually inspected and manually edited, if necessary using BioEdit v7.1.3 (Hall, 1999).

Divergence times were estimated using a relaxed molecular clock approach as implemented in BEAST (Drummond and Rambaut, 2007) with GTR+G+I model selected as the substitution model. A relaxed lognormal molecular clock model was set. The Yule speciation process with a random starting tree was selected as tree prior. The age of the tree root was set to a normal distribution with a mean of – 35 Ma and a standard deviation of 3 (giving a 95% CI ranging from 30.07 – 39.93 Ma) using the age estimate for Vandeae determined by Gustafsson et al. (2010). The prior distribution of the ‘ucld.mean’ parameter was set to an exponential distribution (mean=10.0, initial value=1.0). Four separate runs were performed in BEAST with 50 million generations each, sampling parameters every 1000 generations. Trees were summarized with burn-in values set to the first 25% of trees sampled using TreeAnnotator and were summarized in a maximum clade credibility tree (MCCT).

3.3.2. Diversification analyses

To assess the diversification patterns of plant habitats in Angraecum, we evaluated the state-dependent diversification of this character using the multistate speciation and extinction (MuSSE) models implemented in the R package ‘diversitree’ (FitzJohn, 2012). This habitat character has four potential states (E: canopy epiphyte, I: inselberg lithophyte, L: undergrowth lithophyte, T: trunk epiphyte). Character state distribution was provided in the character

matrix of Andriananjamanantsoa et al. (in chap. 1). Since the MuSSE method requires ultrametric and fully bifurcating trees, we used the MCC tree generated from BEAST as input. All taxa that do not belong to Angraecum were excluded from the analyses using the ‘drop.tip2’ function of the R package ‘phyloch’ (Heibl, 2008).

Six models were tested, one full model allowing all parameters to vary (λ: speciation rate, µ: extinction rate, and q: transition rate), and five constrained models that allow λ and µ to vary while constraining the q parameter. In the second model, only transitions between E and L (E↔L), from E to I (E→I), and from E to T (E→T) were allowed. In the third model, five transitions were allowed (E↔L, E↔T, and E→I). The fourth model allowed five transitions (E↔L, E→T, E→I, and L→T), the fifth model seven (E↔L, E↔T, L↔T, and E→I), and the sixth, eight (E↔L, E↔T, L↔T, T→I, and L→I). After selecting the best fit model (Table 3.1), we set up the prior to be exponential and computed the posterior probabilities of the parameters under a Bayesian framework. The MCMC was run for 10000 generations sampling parameters every 100 generations, and posterior probability distributions were summarized using the function ‘profiles.plot’ implemented in ‘diversitree’ (FitzJohn, 2012).

Table 3.1. Comparison of models tested in MuSSE; bold characters within the table highlight the best fit model selected for analyses. Character states: E, canopy epiphyte; I, inselberg lithophyte; L, undergrowth lithophyte; T, trunk epiphyte; ↔, reversible transition; →, irreversible transition. Abbreviations: AIC, Akaike information criterion; Df, degrees of freedom; lnLik, log likelihood; M, model.

Model transitions allowed Df lnLik AIC M1 All (full model) 20 -300.25 640.50 M2 E↔L, E→I, E→T 12 -321.50 666.99 M3 E↔L, E↔T, E→I 13 -311.11 648.22 M4 E↔L, E→T, E→I, L→T 13 -309.65 645.30 M5 E↔L, E↔T, L↔T, E→I 15 -301.22 632.45 M6 E↔L, E↔T, L↔T, T→I, L→I 16 -307.11 646.22

3.3.3. Biogeographical history analyses

To determine the biogeographical origins of the Angraecinae, ancestral ranges were inferred using the biogeography with Bayesian and likelihood evolutionary analyses using R scripts implemented in the R package ‘BioGeoBEARS’ (Matzke, 2013b). The method implements different models such as the dispersal–extinction–cladogenesis (DEC) (Ree and Smith 2008) or the dispersal–vicariance (DIVA) analysis (Ronquist, 1997) with added parameters (Matzke, 2013a, 2013b). One of its particularities is that it considers founder-event speciation, described as parameter j. Founder-event speciation implies that dispersal to an area outside the ancestral range results in a new species; this is thought to be especially significant in island systems (Moya et al., 1995; Matzke, 2013a; Matzke, 2014). We conducted the analyses using the tree generated from BEAST as input. The following regions were used for the overall analysis: Md, Madagascar; Ms, Mascarenes; Com, Comoros; eA, east Africa; Se, Seychelles and Sri Lanka; As, Southeast Asia; wA, west Africa; and Am, South America. Three models were tested, one unconstrained and two with constraints (M0 and M1). Because of the geographical distance between areas and their geological history, we applied differential dispersal rate parameters between areas in the constrained models (Table 3.2). Since the emergence of archipelagos in the western Indian Ocean occurred from the late Miocene to the Pleistocene (Baksi and Hoffman, 2000; Weyeneth et al. 2008), we designated three time slices related to this temporal framework: 2 Ma as maximum age for the Mascarenes, 7 Ma for the appearance of Mayotte (Comoros), and 15 Ma for the maximal age of the Comoros.

In order to gain a finer comprehension of the biogeographic history of Angraecum, a separate analysis was run in which a greater focus was placed on biogeographic subdivisions within Madagascar and the IOI. We divided Madagascar into five areas according to species distribution taken from Cribb and Hermans (2009), supplemented by our own field data and specimen labels. These areas are congruent with the vegetation zones of Humbert (1955), except that the mountain zone was not treated as a separate area. The west and south were grouped as the west, while the northeast and southeast were treated as separate areas in order to increase the resolution of biogeographic analyses (Fig. 3.1). Geographical units outside the IOI were grouped into a single area because of software limitation in treating more than nine areas. The following regions were retained for biogeographical analyses: within Madagascar:

Table 3.2. Time slices and constraints used in ancestral area reconstructions of subtribe Angraecinae implemented in BioGeoBEARS; age of Mascarenes (2 Ma), Mayotte (7 Ma), and Comoros (15 Ma). Areas within Madagascar: MdN, north; MdNe, northeast; MdSe, southeast; MdC, center; MdW, west. Areas outside Madagascar: Ms, Mascarenes; Cm, Comoros; Se, Seychelles/Sri Lanka; AA, Africa/America/Asia. Abbreviations: M, model constraint; Ma, million years.

Areas MdN MdNe MdSe MdC MdW Ms Cm Se AA time cons- slices traint MdN 1 0.5 0.4 0.3 0.2 0.01 0.3 0.01 0.01 MdNe 0.5 1 0.5 0.5 0.1 0.01 0.3 0.3 0.1 MdSe 0.4 0.5 1 0.5 0.1 0.01 0.01 0.01 0.01 MdC 0.3 0.5 0.5 1 0.5 0.01 0.01 0.3 0.01 MdW 0.2 0.1 0.1 0.5 1 0.01 0.1 0.01 0.01 2 Ma M0, M1 Ms 0.01 0.01 0.01 0.01 0.01 1 0.01 0.01 0.01 Cm 0.3 0.3 0.01 0.01 0.1 0.01 1 0.01 0.01 Se 0.01 0.3 0.01 0.3 0.01 0.01 0.01 1 0.01 AA 0.01 0.1 0.01 0.01 0.01 0.01 0.01 0.01 1 MdN 1 0.5 0.4 0.3 0.2 0.01 0.3 0 0.01 MdNe 0.5 1 0.5 0.5 0.1 0.01 0.3 0 0.1 MdSe 0.4 0.5 1 0.5 0.1 0.01 0.01 0 0.01 MdC 0.3 0.5 0.5 1 0.5 0.01 0.01 0 0.01 MdW 0.2 0.1 0.1 0.5 1 0.01 0.1 0 0.01 7 Ma M1 Ms 0.01 0.01 0.01 0.01 0.01 1 0.01 0 0.01 Cm 0.3 0.3 0.01 0.01 0.1 0.01 1 0 0.01 Se 0 0 0 0 0 0 0 0 0 AA 0.01 0.1 0.01 0.01 0.01 0.01 0.01 0 1 MdN 1 0.5 0.4 0.3 0.2 0.01 0 0 0.01 MdNe 0.5 1 0.5 0.5 0.1 0.01 0 0 0.1 MdSe 0.4 0.5 1 0.5 0.1 0.01 0 0 0.01 MdC 0.3 0.5 0.5 1 0.5 0.01 0 0 0.01 MdW 0.2 0.1 0.1 0.5 1 0.01 0 0 0.01 15 Ma M1 Ms 0.01 0.01 0.01 0.01 0.01 1 0 0 0.01 Cm 0 0 0 0 0 0 0 0 0 Se 0 0 0 0 0 0 0 0 0 AA 0.01 0.1 0.01 0.01 0.01 0.01 0 0 1

MdN, north (Sambirano); MdNe, northeast; MdSe, southeast; MdC, center; E, MdW; outside Madagascar: AA, Africa, America and Asia; Cm, Comoros; Ms, Mascarenes; Se, Seychelles and Sri Lanka.

Fig. 3.1. Biogeographic areas inside Madagascar used in ancestral area reconstructions of subtribe Angraecinae implemented in BioGeoBEARS, colors indicate the different areas and letters denote the 22 administrative regions of Madagascar.

3.4. Results

3.4.1. Divergence time estimates

Phylogenetic relationships and time-calibrated trees used here were presented in chap. 1. The details of the highest posterior density (HPD) for major clades are shown in Table 3.3. The analysis suggested a most recent common ancestor in the late Oligocene (~ 26.1 Ma) for

Angraecinae, and early linage diversification around 14.56 Ma (95% HPD: 9.9 – 19.2 Ma) for Angraecum. Most of the Angraecinae clades diverged from the early Miocene through the Pleistocene. Two waves of diversification were observed in Angraecum, during the Pliocene (~ 6 – 2.6 Ma) and during the Pleistocene (~ 2.6 – 0.2 Ma).

Table 3.3. Ages of crown nodes of relevant clades presented in the phylogeny of Angraecinae inferred from BEAST (from chap. 1; Appendix 8). Comparison between the mean time to most recent common ancestor (MRCA) estimates and the node height highest posterior density (HPD) intervals at 95%.

Angraecinae clades MRCA (node) 95% HPD (node) Outgroup 27.6 [18.2 – 34.8] Angraecinae 26.1 [18.1 – 33.5] Clade I 21 [14.6 – 27.8] Clade II 17.1 [11.7 – 23.3] Clade A 18 [12.1 – 24.5] Clade B 19.4 [13.7 – 25.9] Aeranthes and Jumellea 14.9 [9.2 – 19.9] Aeranthes 10.3 [5.8 – 15.1] Jumellea 7 [4.7 – 11] Angraecum 15.5 [10.4 - 20,6] sect. Pseudojumellea 8.7 [5 – 12.6] sect. Perrierangraecum 9.7 [5.9 – 13.9] sect. Stellariangraecum 5.7 [1.4 – 10.4] sect. Humblotiangraecum 8.5 [5 – 11.9] sect. Hadrangis 1.6 [0.3 – 3.2] sect. Africanae 6.1 [2.7 – 9.7] sect. Angraecoides 2.8 [1 – 4.9] sect. Angraecum 2 [0.5 – 3.6] sect. Arachnangraecum 9.2 [6.1 – 12] sect. Robusta 6.1 [3.1 – 8.7] sect. Boryangraecum 6.3 [3.8 – 8.8]

3.4.2. Habitat diversification

Results from MuSSE showed that model 5 received the best AIC score (Table 3.1), suggesting that the shifts from canopy epiphytes to undergrowth lithophytes or conversely, from canopy epiphytes to trunk epiphytes or conversely, from undergrowth lithophytes to trunk epiphytes or conversely, and from canopy epiphytes to inselberg lithophytes are the most probable habitat transitions among those tested here leading to diversification within Angraecum (Fig. 3.2A). Canopy epiphytes had the highest diversification rate (2.14e-01), followed by trunk lithophytes (2.12e-01) and undergrowth epiphytes (1.99e-01) that had similar rates (Table 3.4). The inselberg lithophytes diversification rate was negative (- 6.90e-01).

Fig. 3.2. (A) Forest strata transition model selected in MuSSE. Arrows indicate direction of transitions; values next to arrows are transition rates; values in bold are diversification rates. E, canopy epiphytes; L, undergrowth lithophytes; T, trunk epiphytes; I, inselberg lithophytes. (B) Speciation rate of each character state.

Table 3.4. Diversification rates of the forest strata character states obtained from the best transition model (5) selected in MuSSE (Table 3.1).

Character states Code speciation (λ) extinction (µ) diversification canopy epiphytes E 2,14e-01 8,95e-08 2,14e-01 undergrowth lithophytes L 1,99e-01 8,78e-08 1,99e-01 trunk epiphytes T 2,12e-01 3,55e-08 2,12e-01 inselberg lithophytes I 7,92e-08 6,90e-01 - 6,90e-01

3.4.3. Ancestral range reconstruction

A comparison of the different models used in the program BioGeoBEARS is presented in Table 3.5. Inclusion of founder-event speciation (DEC+J) in the ancestral area modelling resulted in a better AIC compared to the DEC model. The use of Mascarene and the Comoros Islands’ ages as constraint in DECM1+J appears a logical optimization (species cannot reach the islands before they emerged), but did not yield a significantly better AIC when compared to both the DEC+J and DECM0+J models. A graphic representation of the biogeographical history under DECM1+J model is presented in Figures 3.3. Figure 3.3A illustrates the biogeographic history of tribe Vandeae using eight geographical areas. It suggested a Southeast Asian origin for the tribe, with a westward migration to the WIOR leading to a Malagasy origin of subtribe Angraecinae. Our analyses using finer subdivisions within Madagascar (Fig. 3.3B) yielded congruent results with a more detailed optimization for Angraecinae in Madagascar. It suggested a northeast Malagasy ancestral origin dated to the late Oligocene (~ 26.1 Ma) for subtribe Angraecinae, which started to diverge in the early Miocene (~ 21.1 Ma) to engender two major clades: clade I (African-Malagasy lineages) and clade II (Malagasy lineages). Clade I showed a northeast Malagasy origin and diverged to engender a Malagasy clade A at approximately 18.1 Ma, and an African clade B at approximately 19.5 Ma. Clade B showed E and W Africa as possible ancestral areas. The Malagasy clade II appears to have a northeast Malagasy origin and started to diversify to engender Angraecum at 14.6 Ma, Aeranthes at 10.7 Ma, and Jumellea at 7.8 Ma (Fig. 3.3B).

Table 3.5. Comparison of biogeographic models tested in BioGeoBEARS, bold represent selected model used for interpretation and discussions. Abbreviations: AIC, Akaike information criterion; AICwt, AIC weight; alt, alternative (model using parameter j); null, strandard model; LnL, log likelihood; J, parameter j including founder effect; M: constrained model with differential dispersal rate and time slices; pval, p-values. alt null LnLalt LnLnul pval AIC1 AIC2 AICwt1 AICwt2 DEC+J DEC -707.4 -721.9 7.2e-08 1421 1448 1.00 1.4e-06 DECM0+J DECM0 -586.3 -604.9 1.0e-09 1179 1214 1.00 2.2e-08 DECM1+J DECM1 -645.4 -649.6 0.0036 1297 1303 0.96 0.038

In Angraecum, section Pseudojumellea appears to be the first to diverge within Angraecum. It diversified from the late Miocene through early Pleistocene. Ancestral area reconstruction suggested a northeast Madagascar origin for the section, with an expansion across Madagascar and the IOI during the Pliocene and Pleistocene. Section Stellariangraecum had a northeast origin and migrated west towards the central highland. The large section Perrierangraecum originated from northeast Madagascar and migrated to the center and later to the North. Section Humblotiangraecum, its sister section Hadrangis and section Sobennikoffia originated from the central highland. All sections with an uppermost labellum (Africanae, Angraecum, Arachnangraecum, Boryangraecum, Nana, and Oeoniella) originated from the northeast and colonized large ranges from Eastern Africa to WIOR and Sri Lanka.

Fig. 3.3. Ancestral area reconstructions of subtribe Angraecinae from DECM1+J analysis implemented in BioGeoBEARS: A, reconstruction using eight geographic areas (more detailed outside Madagascar); B, reconstruction using nine geographic areas (more detailed within Madagascar); dash lines indicate time slices. Areas within Madagascar: MdC, center; MdN, north; MdNe, northeast; MdSe, southeast; MdW, west. Areas outside Madagascar: AA, Africa/America/Asia; Am, South America; As, Southeast Asia; eA, east Africa; wA, west Africa; Cm, Comoros; Ms, Mascarenes; Se, Seychelles/Sri Lanka.

MdMd Angraecum.tenuispica Fig. 3.3A. part 1 Md MdMdMd Angraecum.pauciramosum MdMd Md Angraecum.musculiferum Md MdMd Md Angraecum.cf.pauciramosum MdMd Md Md Angraecum.caricifolium Md MdMd Md Angraecum.madagascariense Md Md Ms Angraecum.tenuifolium Md Md Angraecum.hermannii MdMd Md MdMs Angraecum.cordemoyi Md MdMd Angraecum.pinifolium Md Md MdMd MdMd Angraecum.cf.pinifolium eA Angraecum.sacciferum MdMdMdMd eA Md Md Angraecum.baronii Md Ms Angraecum.obversifolium Md Md Angraecum.cf.sacculatum Md MdMd Md Md Md Md Ms Angraecum.costatum Md Md Angraecum.ochraceum Md Md Md Md Md Md Angraecum.chaetopodium Md Md Md Md Angraecum.setipes SeSe Angraecum.cf.zeylanicum MdMdMsCmSe Md Md MdMsCmMdMsCm Angraecum.caulescens MdMd Md Md Md Md Angraecum.huntleyoides MdMd Md MdMsCm Angraecum.calceolus Md MdMsCm Angraecum.multiflorum Md Md Angraecum.amplexicaule Md Md MdMd Angraecum.serpens Md Md Md Md MdMd Angraecum.moratii Md Md Md MdCm Angraecum.florulentum Angraecum.penzigianum MdMdMd Md Md Md Md Md Angraecum.oblongifolium MdMd Angraecum.eburneum.var.longica MdMd MdMdMdMsCmMd Angraecum.eburneum.var.superbu Md MdMd Md Angraecum.eburneum.var.xerophi MdMdMsCm Angraecum.eburneum.var.eburneu MsMs Angraecum.corrugatum Md Md Md MdMd MdMd Angraecum.conchoglossum Angraecum.linearifolium MdMdMd Md MsMs Angraecum.expansum eA eA Md Md eAeA Angraecum.conchiferum Md Md Angraecum.mirabile Md Md Md Md Angraecum.cf.germinyanum Md Md MdMd Md MdMd Md Angraecum.arachnites MdMd Md Md Md Md Angraecum.germinyanum Md Ms Angraecum.appendiculatum Cm Cm Angraecum.scottianum Md Md Md Md Angraecum.pseudofilicornu MdMd Md Md Angraecum.polystachyum Md Md Angraecum.triangulifolium Md Md Md Md Md Md Angraecum.sedifolium Md Md Md MdMs Angraecum.pingue Md Md Md Angraecum.nanum eA eA Angraecum.teres eA eA Md Md eA eA Angraecum.dives Md Md Angraecum.compactum MdMd Md MdMd Md Angraecum.clareae Angraecum.leonis Md Md Md Md Md Md Angraecum.viguieri Md Md Md Md Angraecum.magdalenae Md Md MsMs Angraecum.striatum Ms Ms Md MdMsMsMs Angraecum.bracteosum Md Md Ms Ms Angraecum.cadetii Md Md Angraecum.humbertianum Md Md Angraecum.pseudodidieri MdMd MdMdMd Md Angraecum.ankeranense MdMdMd Md Angraecum.didieri Md Md Angraecum.elephantinum Md Md Md Ms Ms Angraecum.liliodorum MdMd MdMdMd Md Angraecum.cf.liliodorum Md Md Md Angraecum.obesum Md Md Angraecum.letouzeyi MdMd MdMdMd MdMs Angraecum.cornigerum Md Md Angraecum.sterophyllum MdMd Md MdMd Md Angraecum.cf.rutenbergianum Md Md Md Angraecum.cf.elephantinum MdMd Angraecum.peyrotii Md Md MdMd MdMd Angraecum.cf.peyrotii Md MdMd Md Md Md Angraecum.cf.breve MdMd Md MdMs Angraecum.borbonicum Md Md Md Md Angraecum.bicallosum Md Md Angraecum.lecomtei Md Md Md Md Md Md Angraecum.drouhardii Md Ms Angraecum.cucullatum Md Md Angraecum.praestans MdMd Md Md Angraecum.sororium Md Md Md Md Angraecum.sesquipedale Md MdMs Angraecum.pectinatum MdMd Md Md Angraecum.dasycarpum MdMd MdMdMsCm Angraecum.mauritianum Md Md MdMd Md Angraecum.danguyanum Md Md Md Md MdMd MdMs Angraecum.ramosum Md Md Md Md Angraecum.cf.panicifolium Md Md Md Md Angraecum.filicornu MdMd Angraecum.panicifolium Md Md MdMd Angraecum.cf.humblotianum Md Md Lemurorchis.madagascariensis

35 30 25 20 15 10 50Millions of years ago

Md Md Jumelleajspathulata MdMd Md MdMd Md Jumelleajhyalina Jumelleajlongivaginans Figj 3j3Aj part 2 MdMd Md Md Md Md Jumelleajdensifoliata Md Md Md Md MdMdCm Jumelleajarachnantha Md Jumelleajjumelleana MdMd Md Md Md Md Md Jumelleajbrevifolia MdMd MdMdMd MdCm Cm Jumelleajanjouanensis Md Md Jumelleajmajalis Md Md Jumelleajteretifolia Md Md Jumelleajrigida Md Md Md Jumelleajmaxillarioides Md MdMd Md Md Md Md Jumelleajtenuibracteata Md Md Md Jumelleajbosseri Md Md Jumelleajarborescens Md MdMd Md Md Md Md Md Jumelleajfrancoisii Md Md Jumelleajexilis Md Md Aeranthesjlongipes MdMd Md Md MdMd Md Aeranthesjcaudata MdMd Md Aeranthesjstrangulata Md Md Aeranthesjneoperrieri MdMd MdMd Aeranthesjgrandiflora MdMd Md MdMd MdMdMd Aeranthesjarachnites MdMd Aeranthesjvirginalis Md Md Md Md Aeranthesjmoratii Md Md Md Md Aeranthesjaemula Md Md Aeranthesjtenella wAwA Pectinariellajpungens wAwA wA wAwAwA Pectinariellajatlantica wA wA wA wA Pectinariellajgabonensis wA wA wA wA Pectinariellajdoratophylla wA wA wA wA Pectinariellajsubulata eA wA wA Listrostachysjpertusa eA eA Tridactylejfilifolia eA eA eA eA Tridactylejbicaudata eAeA eA eA Ypsilopusjviridiflorus eA eA eA eA eA eA Ypsilopusjlongifolius eA Rangaerisjamaniensis eA eAeA eA eA eA eAwA Rangaerisjmuscicola eAwA eAwA Cyrtorchisjchailluana eA eAwA Cyrtorchisjarcuata eAeA eA eA eAwA eAwA eA eAwA Podangisjdactyloceras wA wA Ancistrorhynchusjrecurvus Md Md wA wA wA wA wA wA Ancistrorhynchusjcapitatus wA eAwAAm Calyptrochilumjchristyanum eA eA eA eA Cribbiajconfusa eA eA eA eA eA eA Cribbiajbrachyceras eA eA eA eA Rhipidoglossumjsubsimplex eA eA Rhipidoglossumjxanthopollinium eA eA eA eA eA Mystacidiumjflanaganii eA eA eA eA eA Mystacidiumjcapense eA eA eA eA Sphyrarhynchusjschliebenii eAeA eAeA eA MdMsCmeAwA Angraecopsisjparviflora eA eA Microcoeliajstolzii eA MdeA Md Md Microcoeliajgilpinae eA eA Solenangisjwakefieldii eA eAwA wA wA Solenangisjclavata Cm Cm Aerangisjhildebrandtii Cm Cm MdMd Cm Cm Aerangisjhariotiana MdeAwAMdeAwA MdMd Md Aerangisjellisii Md MdMd Md Aerangisjmacrocentra MdwAMdwA Md Md Aerangisjpunctata wA wA Eurychonejrothschildiana wA wA wA wA wA Eurychonejgaleandrae eA eAwA Bolusiellajmaudiae eA eAwA eA CmeA Bolusiellajiridifolia wA wA wAwA Dolabrifoliajdisticha wA wA wA wA wAwA Dolabrifoliajbancoensis eAMdeAwA wA wA wA wA Dolabrifoliajcfjaporoides wA wA Dolabrifoliajpodochiloides MdMsCmeAAswA wA wA Angraecoidesjerecta wA wA wA wA Angraecoidesjcultriformis wA wA wAwA Angraecoidesjchevalieri wA wA wA wA wAwA Angraecoidesjcfjmoandense Am Am Dendrophylaxjfunalis Am Am wA wA Am Am Campylocentrumjfasciola wAwA Eichlerangraecumjinfundibulare wA wA wAwAwA Eichlerangraecumjbirrimense wA wA wAwA Eichlerangraecumjcfjeichlerian wA eAwA Diaphananthejvesicata Md MdeAwA wA wA wA wA wA eAwA Diaphananthejsarcophylla wA wA Diaphananthejpellucida Md Md Cryptopusjpaniculatus Md Md Md Md Md MdMs Cryptopusjelatus Oeoniajrosea Md Md Md Md Md Md Neobathieajgrandidierana Md Md Md Md Md Md Acauliajrhynchoglossa Md Md Lemurellajpapillosa Md Md Md Md Md Md Lemurellajpallidiflora Md Md Parangraecumjpterophyllum Md Md Md Md Md Md Md MdeA Parangraecumjcfjhumile Md Md Parangraecumjperparvulum Md Md Beclardiajmacrostachya Md Md Md Md Erasanthejhenricii As As Vandajtricolor As As As As As As Aeridesjodorata As As As As Acampejochracea MdMsCmeAAswA MdMsCmeAAswA As As Phalaenopsisjcornucervi MdMsCmeAwA MdMsCmeAwA Polystachyajfulvilabia Eocene Oligocene Miocene Pliocene-Pleistocene

MdNeMdNeMdC Angraecum.tenuispica MdNe Fig. 3.3B. part 1 MdNeMdNeMdNeMdC Angraecum.pauciramosum MdNMdNeMdCMdNeMdNe Angraecum.musculiferum MdNeMdNeMdCMdNeMdNeMdC Angraecum.cf.pauciramosum MdNeMdNeMdNMdNeMdCMdNe Angraecum.caricifolium MdNeMdNeMdNe MdNeMdC Angraecum.madagascariense MdNe MdNe Ms Angraecum.tenuifolium MdNe Ms Angraecum.hermannii MdNeMdNe MdNe MdNeMdCMs Angraecum.cordemoyi MdNe MdNeMdNe Angraecum.pinifolium MdNe MdNe MdNe MdNeMdNe Angraecum.cf.pinifolium AA AA Angraecum.sacciferum MdNeMdNeMdNeMdNe MdC MdNMdC Angraecum.baronii MdNe Ms Angraecum.obversifolium MdNe MdNe Angraecum.cf.sacculatum MdNeMdNeMdNeMdNe MdNeMdNeMdC MdNe Ms Angraecum.costatum MdNe MdNMdNe Angraecum.ochraceum MdNe MdNe MdNeMdNe MdNe MdNMdNe Angraecum.chaetopodium MdNeMdNe MdNe MdNeMdSeMdCAngraecum.setipes SeSe Angraecum.cf.zeylanicum MdNMdCCmSeMdNMdCCmMsSe MdNMdNeMdCCmSeMdNe MdNMdCCmMsMdNMdCCmMs Angraecum.caulescens MdNeMdNe MdNeMdNe MdNe MdNe Angraecum.huntleyoides MdNeMdNeMdNeMdNMdNeMdSeMdCMdWCmMsAngraecum.calceolus MdN MdNMdCCmMs Angraecum.multiflorum MdNe MdNe Angraecum.amplexicaule MdNe MdNe MdNeMdNe Angraecum.serpens MdNeMdNe MdNeMdNeMdNMdNeMdNe Angraecum.moratii MdNe MdNeMdNe MdNeCm Angraecum.florulentum Angraecum.penzigianum MdNeMdNe MdNeMdCMdNe MdNeMdC MdC MdC Angraecum.oblongifolium MdCMdC Angraecum.eburneum.var.longica MdNMdNeMdSeMdCMdWCmMsMdNMdNeMdSeMdCMdWCm MdNMdNeMdSeMdCMdWCmMdNMdNeMdSeMdCMdWCmMdNMdNeMdSeMdWCmMsMdNMdNeMdSeMdWCmMsAngraecum.eburneum.var.superbu MdNe MdNMdNeMdSeMdCMdWCmMdSeMdSeMdW Angraecum.eburneum.var.xerophi MdNMdNeMdSeCmMsMdNe Angraecum.eburneum.var.eburneu MsMs Angraecum.corrugatum MdNeMdNMdNeMdCMs MdNe MdNeMdNeMdNMdNeMdCMdNMdNeMdC Angraecum.conchoglossum Angraecum.linearifolium MdNeMdNeMdNe MdNMdNeMdC MsMs Angraecum.expansum AA AA MdNe MdNe AAAA Angraecum.conchiferum MdNe MdNMdNeMdC Angraecum.mirabile MdNe MdNe MdNeMdNe Angraecum.cf.germinyanum MdNeMdNe MdNeMdNe MdNe MdNeMdNMdNeMdCMdNe Angraecum.arachnites MdNeMdNe MdNe MdNe MdNMdNeMdCMdNe Angraecum.germinyanum MdNe Ms Angraecum.appendiculatum Cm Cm Angraecum.scottianum MdNe MdNe MdNe MdNMdNeMdC Angraecum.pseudofilicornu MdNeMdNe MdNe MdNMdNe Angraecum.polystachyum MdCMdNeMdC Angraecum.triangulifolium MdNeMdCMdNeMdC MdC MdNMdNeMdCMdCMdNeMdC Angraecum.sedifolium MdC MdC MdN MdNMs Angraecum.pingue MdC MdC MdC Angraecum.nanum AA AA Angraecum.teres AA AA MdNe MdC AA AA Angraecum.dives MdCMdNMdNeMdSeMdCAngraecum.compactum MdCMdC MdC MdCMdC MdC Angraecum.clareae Angraecum.leonis MdC MdC MdC MdNMdCMdW MdC MdNMdNeMdC Angraecum.viguieri MdC MdC MdC MdC Angraecum.magdalenae MdC MdC MsMs Angraecum.striatum Ms Ms MdC MdCMsMsMs Angraecum.bracteosum MdCMdC Ms Ms Angraecum.cadetii MdC MdSeMdCMdW Angraecum.humbertianum MdNMdN Angraecum.pseudodidieri MdNMdN MdNMdNMdNMdN Angraecum.ankeranense MdCMdCMdNMdNMdNe Angraecum.didieri MdC MdC Angraecum.elephantinum MdNe MdCMdC Ms Ms Angraecum.liliodorum MdCMdC MdCMdC Angraecum.cf.liliodorum MdCMdCMdC MdCMdC MdC Angraecum.obesum MdC MdNeMdC Angraecum.letouzeyi MdCMdC MdCMdCMdC MdCMs Angraecum.cornigerum MdCMdC Angraecum.sterophyllum MdCMdC MdCMdCMdCMdC Angraecum.cf.rutenbergianum MdC MdC MdC Angraecum.cf.elephantinum MdCMdC Angraecum.peyrotii MdC MdC MdNeMdNe MdCMdC Angraecum.cf.peyrotii MdNeMdNeMdCMdC MdC MdNeMdNMdNe Angraecum.cf.breve MdCMdC MdC MdCMs Angraecum.borbonicum MdNe MdNe MdNe MdNMdNe Angraecum.bicallosum MdNe MdNe Angraecum.lecomtei MdNe MdNe MdNe MdNe MdNeMdNe MdNMdNe Angraecum.drouhardii MdNe Ms Angraecum.cucullatum MdN MdNMdSeMdW Angraecum.praestans MdNeMdNe MdSe MdNMdSeMdC Angraecum.sororium MdNe MdNMdNeMdSe MdNe MdNMdNeMdSeAngraecum.sesquipedale MdNeMdNMdNeMdCMsAngraecum.pectinatum MdNeMdNe MdNe MdNMdNe Angraecum.dasycarpum MdNeMdNe MdNMdNeMdSeMdCCmMsMdNe Angraecum.mauritianum MdNeMdNe MdNeMdNe MdNeMdNeMdC Angraecum.danguyanum MdNeMdNe MdNeMdNeMdNe Ms Angraecum.ramosum MdNe MdNeMdNe MdNe Angraecum.cf.panicifolium MdNe MdNe MdNe MdNeMdSeMdCAngraecum.filicornu MdNeMdCMdNe Angraecum.panicifolium MdNe MdNMdNeMdC MdNMdNeMdCMdNMdNeMdC Angraecum.cf.humblotianum MdNe MdNeMdC Lemurorchis.madagascariensis

35 30 25 20 15 10 50Millions of years ago

MdCMdC Jumellea.spathulata MdCMdC MdCMdNMdNeMdCMdCMdC Jumellea.hyalina Jumellea.longivaginans Fig. 3.3B. part 2 MdCMdC MdNeMdNMdNe MdC MdC Jumellea.densifoliata MdC MdCCm MdNe MdC Cm Cm Jumellea.arachnantha MdC MdNeMdC Jumellea.jumelleana MdNeMdNe MdNe MdNe MdSe MdSe Jumellea.brevifolia MdSeCmMdSeCm MdNeMdNeMdNe MdNeMdSeCmCm Cm Jumellea.anjouanensis MdNeMdNMdNe Jumellea.majalis MdNe MdNeMdC Jumellea.teretifolia MdC MdC Jumellea.rigida MdNe MdCMdC Jumellea.maxillarioides MdNeMdCMdCMdCMdC MdNeMdC MdN MdN Jumellea.tenuibracteata MdC MdC Ms Jumellea.bosseri MdC MdC Jumellea.arborescens MdCMdCMdC MdC MdC MdNeMdC MdC MdNeMdC Jumellea.francoisii MdC MdCMs Jumellea.exilis MdCMdCMs Aeranthes.longipes MdNe MdC Aeranthes.caudata MdNe MdNMdNeMdCCmMdCMdC MdNeMdC MdCMs Aeranthes.strangulata MdNe MdNe Aeranthes.neoperrieri MdNMdNeMdNeMdNMdNeMdSeMdNMdNeMdSe Aeranthes.grandiflora MdNMdNeMdSeMsMdNMdNeMdSeMsAeranthes.arachnites MdN MdNMdNeMdSeCmMdNMdNeMdNMdNeMdSeCmMsMsMs CmCm Aeranthes.virginalis MdNe MdNe MdN MdN Aeranthes.moratii MdNe MdNe MdNe MdNMdNeMdC Aeranthes.aemula MdNe MdNMdNe Aeranthes.tenella AAAA Pectinariella.pungens AAAA AA AAAAAA Pectinariella.atlantica AA AA AA AA Pectinariella.gabonensis AA AA AA AA Pectinariella.doratophylla AA AA AA AA Pectinariella.subulata AA AA AA Listrostachys.pertusa AA AA Tridactyle.filifolia AA AA AA AA Tridactyle.bicaudata AAAA AA AA Ypsilopus.viridiflorus AA AA AA AA AA AA Ypsilopus.longifolius AA AA Rangaeris.amaniensis AA AAAA AA AA AA Rangaeris.muscicola AA AA Cyrtorchis.chailluana AA AA Cyrtorchis.arcuata AAAA AA AA AA AA AA AA Podangis.dactyloceras AA AA Ancistrorhynchus.recurvus MdNe MdNe AA AA AA AA AA AA Ancistrorhynchus.capitatus AA AA Calyptrochilum.christyanum AA AA AA AA Cribbia.confusa AA AA AA AA AA AA Cribbia.brachyceras AA AA AA AA Rhipidoglossum.subsimplex AA AA Rhipidoglossum.xanthopollinium AAAA AA AA AA AA Mystacidium.flanaganii AA AA AA AA Mystacidium.capense AA AA AA AA Sphyrarhynchus.schliebenii AAAA AAAA AA MdNMdNeAACmMsAngraecopsis.parviflora AA AA Microcoelia.stolzii AA AA AA MdNMdNeMdSeMdCMicrocoelia.gilpinae AA AA Solenangis.wakefieldii AA AA AA AA Solenangis.clavata Cm Cm Aerangis.hildebrandtii Cm Cm MdNeMdNe Cm Cm Aerangis.hariotiana AA AA MdNeMdNe MdNMdNeMdSeMdCAerangis.ellisii AA MdNeMdNe MdNMdNeMdC Aerangis.macrocentra AA AA MdNe MdNeMdC Aerangis.punctata AA AA Eurychone.rothschildiana AA AA AA AA AA Eurychone.galeandrae AA AA Bolusiella.maudiae AA AA AA AACm Bolusiella.iridifolia AA AA AAAA Dolabrifolia.disticha AA AA AA AA AAAA Dolabrifolia.bancoensis AA AA AA AA AA AA Dolabrifolia.cf.aporoides AA AA Dolabrifolia.podochiloides MdNeAACmMs AA AA Angraecoides.erecta AA AA AA AA Angraecoides.cultriformis AA AA AAAA Angraecoides.chevalieri AA AA AA AA AAAA Angraecoides.cf.moandense AA AA Dendrophylax.funalis AA AA AA AA AA AA Campylocentrum.fasciola AAAA Eichlerangraecum.infundibulare AAAA AA AAAAAA Eichlerangraecum.birrimense AA AA AAAA Eichlerangraecum.cf.eichlerian AA AA Diaphananthe.vesicata MdNe MdNeAA AA AA AA AA AA AA Diaphananthe.sarcophylla AA AA Diaphananthe.pellucida MdNe MdNe Cryptopus.paniculatus MdNe MdNe MdNe MdNe MdNe Ms Cryptopus.elatus Oeonia.rosea MdNe MdNe MdNe MdNeMdSeMdC MdC MdNMdSeMdC Neobathiea.grandidierana MdNe MdC MdNeMdNe MdNe MdNeMdC Acaulia.rhynchoglossa MdNe MdNMdNe Lemurella.papillosa MdNe MdNe MdNe MdNe MdNe MdNMdNe Lemurella.pallidiflora MdNe MdNeMdC Parangraecum.pterophyllum MdNe MdNe MdNe MdNe MdNe MdNe MdNe MdNeMdCAA Parangraecum.cf.humile MdNe MdNeMdC Parangraecum.perparvulum MdN MdNMdNeMdSeMdCMsBeclardia.macrostachya MdN MdN MdN MdNMdC Erasanthe.henricii AA AA Vanda.tricolor AA AA AA AA AA AA Aerides.odorata AA AA AA AA Acampe.ochracea AACmMs AACmMs AA AA Phalaenopsis.cornucervi AACmMs AACmMs Polystachya.fulvilabia Eocene Oligocene Miocene Pliocene-Pleistocene

3.5. Discussion

3.5.1. Stratification changes and orchid diversification

Colonization of new habitats is known to drive species diversification in response to ecological opportunities (Reddy et al., 2012). The evolution of epiphytism, often associated with CAM photosynthesis, has been shown to be a key innovation leading to species diversification in the Orchidaceae (Givnish et al., 2015). Habitat change also has been demonstrated to be at the origin of diversification in subfamily Epidendroideae (Freudenstein and Chase, 2015), to which subtribe Angraecinae belongs. Angraecum is symplesiomorphically epiphytic, with over 95% of the species occupying that particular habitat. The shift from canopy epiphytes to undergrowth lithophytes and from canopy epiphytes to trunk epiphytes resulted in similar diversification rates in Angraecum (Fig. 3.2). Such shifts in the stratum occupied within the forest could be construed as changes in habitat and thus new habitat colonization events. The negative diversification rate for inselberg lithophytes could be the result of slow extinction (Donoghue and Sanderson, 2015) or an absence of speciation. These types of habitats are usually isolated and the species occupying them are highly specialized (e.g. A. eburneum var. longicalcar, A. praestans, A. sororium).

The colonization of new forest strata may expose the plants to new pollinator types, providing new evolutionary opportunities. It is possible that changes in strata may be associated with pollinator shifts. We have noticed during our field surveys that all species with green flowers (e.g. sections Boryangraecum) occupy exclusively the understory layer (trunk epiphytes), indicating a possible shift in pollinator. This could limit pollinator competition with species from other strata. To our knowledge, this is the first study to report the effect of changes in the strata occupied on orchid diversification and our results show that all strata appear to have contributed similarly to the diversification of Angraecum. Further studies are required (especially in Madagascar) to determine the pollinators of species adapted to canopy, trunk, or rock substrates.

3.5.2. Malagasy origin of subtribe Angraecinae

Our results suggest a Malagasy origin for subtribe Angraecinae in the early Miocene, with a probable source from sotheast Asia (Fig. 3.4). Using the ancestral area of Bremer (1992) to study the historical biogeography of Mascarene angraecoid orchids, Micheneau et al. (2008a) showed that Madagascar was potentially the ancestral area of the Angraecinae. Two major clades are observed in the Angraecinae, an African-Malagasy clade I with a Malagasy origin (Fig. 3.3A) that diverged in the late Miocene (~ 21 Ma), giving rise to mostly small Malagasy (clade A) and African (clade B) genera, and a Malagasy clade II that diverged later in the mid- Miocene (~ 17 Ma) and diversified into three large genera, Angraecum (~ 14 Ma), and Aeranthes (~ 10 Ma) sister to Jumellea (~ 7 Ma). Micheneau et al. (2008a) suggested an African origin for clade I, while our results support a Malagasy origin (Appendix 10). The Malagasy clade A has a northeast origin (Fig. 3.3B) and colonized the rest of the Island, except the western part. Northeast here refers to the northern half of the lowland tropical forests (eastern forests 0 – 800 m) of Humbert (1955). Two independent migrations, originating from northeast Madagascar are observed, one to east Africa during the Pliocene (~ 6.1 Ma) and one to the Mascarenes during the Pleistocene (~ 3.1 Ma). Several cases of out of Madagascar dispersal events have been proposed recently (e.g. Raxworthy et al., 2002; Stankiewicz et al., 2006; Krüger et al., 2012; Buerki et al., 2013), supporting our findings. Within the African Angraecinae clade B, no specific area has been found as a possible site of origin within Africa (east or west), putatively due to the limited sampling in our analysis. An increased sampling may help resolve the distribution history of Angraecinae within continental Africa. Nonetheless, a dispersal event that originated from western Africa during the late Miocene gave rise to the American Angraecinae (Campylocentrum and Dendrophylax) group. Furthermore, at least four secondary dispersals (essentially in Aerangis, Angraecopsis, Bolusiella, and Microcoelia) from continental Africa back to Madagascar and to the WIOR were observed during the Pliocene and Pleistocene (Fig. 3.3A; Fig. 3.4).

Fig. 3.4. Movements of dispersal in subtribe Angraecinae, arrows indicate the direction of the migration and numbers designate the age estimate of the events in million years. Red arrow indicates the early migration of subtribe Angraecinae from Southeast Asia towards Madagascar and Africa; blue arrows indicate the migration path of Angraecinae from Madagascar towards Africa then South America; green arrow denotes secondary migrations from Africa to Madagascar.

3.5.3. Origin of Angraecum and its dispersal within Madagascar

Our results suggest a Malagasy origin of Angraecum in the mid-Miocene. Two other Malagasy Angraecinae genera, Aeranthes and Jumellea, diversified as well during this period (Fig. 3.3B). The Miocene was characterized by the northern migration of Madagascar towards the equator, the establishment of the monsoon in the WIOR (Buerki et al., 2013), and the expansion of the Malagasy tropical forest, particularly its northern part (Yoder and Nowak, 2006; Buerki et al., 2013), which corresponds to our northeast region. These events were highlighted as the main diversification factors in Angraecum (chap. 1). The colonization of Angraecum within Madagascar started during the mid-Miocene and continued into the Pleistocene. The genus started to expand from northeast Madagascar (the northern half of the lowland humid tropical forest) and migrated to the central highland (highland tropical forest

adjacent to the lowland forest of the northeast) during the late Miocene-early Pliocene. As in Angraecum, the northern tropical rain forest of Madagascar has been considered to be the point of origin of the expansion of many recent Malagasy taxa (Raxworthy and Nussbaum, 1995; Vences et al., 2009; Anthony et al., 2010; Rakotoarinivo et al., 2013). Two migrations to the West (A. humbertianum and A. praestans) originated from the center and the northeast, respectively, and occurred independently during the late Miocene (Fig. 3.3B). A great number of dispersal events within Madagascar happened independently during the Pliocene- Pleistocene, most of them originating from the northeast or the center. A similar pattern of dispersal is also observed in Aeranthes and Jumellea (Fig. 3.3B), though our data are incomplete there.

3.5.4. Out of Madagascar dispersal of Angraecum

Dispersal out of Madagascar happened during the late Miocene-early Pliocene to east Africa, during the early Pliocene to the Comoros, and during the Pleistocene to the Mascarenes, Seychelles and Sri Lanka (Fig. 3.5). In most cases, dispersal originated from the northeast or the center of Madagascar, and involved species that are still present on the island. A single speciation event in east Africa, leading to sect. Africanae (A. divers and A. teres), originated from central Madagascar during the late Miocene (Fig. 3.3B). At least two dispersal events (A. conchiferum, A. sacciferum) took place independently from northeast Madagascar to east Africa during the Pliocene. Additionally, two further east African species not included in our analyses probably originated from the center (A. stella-africae, a member of sect. Perrierangraecum related to A. rutenbergianum) and the northeast of Madagascar (A. eburneum var. giryamae, section Angraecum). The absence of Angraecum sensu stricto lineages in west Africa could be explained by the existence of a natural barrier between the two regions. Couvreur et al. (2008) reported that the Guineo-Congolian and east African regions are geographically isolated by a ca. 1000 km-wide north-south arid corridor, creating an effective barrier to dispersal for rainforest taxa. The isolation of the east African rain forests happened essentially during the Oligocene-Early Miocene (Couvreur et al., 2008) before the dispersion of Angraecum. Because uplift of the two archipelagos occurred in the late Miocene

for the Comoros (Weyeneth et al. 2008) and in the Pliocene for the Mascarenes (Baksi and Hoffman, 2000), colonization of these islands could not have happened before these dates, which were used as time slice constraints in our analyses (Table 3.2). Dispersal to the Comoros happened several times starting in the early Pliocene and the majority during the Pleistocene (Fig. 3.3B). Most of the lineages found there originated from northeast Madagascar, and only a few endemic species are recorded (Govaerts et al., 2015). The single dispersal event observed to the Seychelles archipelago and Sri Lanka (A. cf. zeylanicum) may be a secondary migration from the Comoros that was reached first in the Pleistocene (Fig. 3.3B). Our results show that dispersal events from Madagascar to the Mascarenes occurred several times independently (Fig. 3.3B). The lineages came essentially from the northeast or the central highland of Madagascar during the Pleistocene. One dispersal event originating from the central highland of Madagascar during the Pleistocene, gave rise to the small radiation of the endemic section Hadrangis, characterized by the adoption of unusual pollination modes in Angraecum (Micheneau, 2005; Micheneau et al., 2009).

Fig. 3.5. Out of Madagascar dispersal of Angraecum, arrows indicate the direction of the migration and numbers designate the age estimate of the events in million years.

3.5.5. Dispersal mechanisms and colonization in Angraecum

The young age of the Angraecinae, approximately 26 Ma (chap. 1), cannot support hypotheses of vicariance, which would require a Cretaceous age in Madagascar, suggesting that long-distance dispersal was the main mechanism to explain area establishment in the group. In Angraecum, migration within or out of Madagascar could be explained by two major phenomena, the frequent passage of cyclones and prevailing winds in the region. Cyclones are accidental disturbances affecting each year the territory of Madagascar and the western Indian Ocean region, especially in summer (December to April). The average number of cyclones passing through Madagascar is estimated at 10 per season (Stankiewicz et al., 2006). Their origin lies in the clash of two drafts on the opposite front of the trade winds and the monsoon (Joly, 1941). These clashes occur only during the summer because in winter monsoon follows the same direction as the trade winds (Joly, 1941). During the cyclonic season, the wind simultaneously sweeps out several areas, shuffling the air in the entire region on its way. Cyclones are formed generally in the Indian Ocean, rarely in the Mozambique Channel. They go westward by the Mascarene Islands, pass the east coast of Madagascar, and continue northward to the Comoros islands or southward to end up in the Mozambique Channel. A return to the Indian Ocean is often observed. However, the storms rarely go eastwards over the Mozambique Channel, and none of them goes over Africa and returns to Madagascar (Stankiewicz et al., 2006). From the Early Miocene (~ 23 Ma), ocean currents and prevailing winds flowed from the Indian Ocean towards Africa, which was not the case during the Eocene (Rabinowitz and Stephen, 2006; Ali and Huber, 2010; Weyeneth et al., 2011).

Oceanic islands are known for their species diversity and high endemism (Goldberg et al., 2014, Goodman et al., 2015, Lewis et al., 2015), and isolation is considered to be the main reason (Weigelt and Kreft, 2013; Mallet et al., 2014). The two Indian Ocean archipelagos appear to be an expansion area for Angraecum, with approximately 45 species and 15 species recorded (Govaerts et al., 2015) from the Mascarenes and the Comoros, respectively. The uplift of these volcanic islands opened up new niches for colonization (Buerki et al. 2013), potentially driving diversification. Dispersal probability is expected to decrease with distance so that gene flow is generally lower between geographically distant populations causing a pattern of isolation-by-distance (Mallet et al., 2014). This could explain the diversification and

high endemicity of the Angraecum species found in the Mascarenes (with one endemic section, Hadrangis) compared to other areas, essentially Africa and Comoros, which are half as far to Madagascar and harbour less species diversity. The differentiation of the new endemic lineages in the Mascarenes, essentially section Hadrangis, could be viewed as a founder-event speciation specific to this island. Jacquemyn et al. (2005) demonstrated the significant role of elevational gradients on the diversification of orchids in Reunion. Acharya et al. (2011) came to the same conclusion when studying Himalayan orchids. This could be explained by ecological similarities between Madagascar and Mascarene regions (Buerki et al. 2013). However, it has been reported that the Mascarenes have a greater affinity to the Malagasy biodiversity than do the Comoros (Buerki et al. 2013), suggesting that cyclones may have been more effective than prevailing winds in Angraecum dispersal.

3.6. Conclusion

The study of the historical biogeography of Angraecinae provides valuable insights into migration and colonization events in the Western Indian Ocean Region and Africa. Our results suggest an out of Madagascar dispersal for the subtribe during the Pliocene- Pleistocene. The origin of Angraecum is in Madagascar, more precisely the northeast of the island where the lowland wet tropical forest occurrs. Local colonization from the northeast to the central highland happened first during the late Miocene, followed by an expansion to the north and the southeast during the Pliocene-Pleistocene. Long-distance dispersal from the northeast or the central highland of Madagascar came later during the Pliocene-Pleistocene into two opposite directions: eastward to the Mascarenes and westward to the Comoros and Africa. The existence of natural barrier between the east and west tropical African forest may explain why Angraecum is restricted to east Africa, Comoros, Madagascar, Mascarene, Seychelles and Sri Lanka. Two main dispersal factors are suggested: cyclonic events and prevailing winds (trade wind and monsoon). Our results suggest that cyclones are likely more effective in orchid dispersal than prevailing winds. Nonetheless, the colonization and diversification of Angraecum in Archipelagos (Comoros and Mascarenes) reflects the importance of new habitat and niche similarities in species expansion. Changes of habitat

100

(stratification) had no significant effect on diversification patterns in Angraecum. However, this change could be correlated with a pollinator shift, which requires further investigation to determine the mechanism undergone between strata.

3.7. Acknowledgements

We are very thankful to S. Joly (Montreal Botanical Garden), F.J. Lapointe, E. Gagnon (University of Montreal) and S. Renaut (Quebec Centre for Biodiversity Science) for their comments, assistance and expertise. This study was supported by Omaha’s Henry Doorly Zoo and Aquarium, and the Madagascar Biodiversity Partnership.

101

Conclusion

Notre étude du genre Angraecum à Madagascar a permis d’évaluer plusieurs hypothèses, notamment la monophylie du genre Angraecum sensu stricto et l’exclusion des sections africaines sensu Garay. Toutefois, nous avons reconnu une nouvelle section endémique de l’Afrique de l’est (sect. Africanae). Nous avons aussi démontré que les caractères morphologiques sont impliqués dans la diversification d’Angraecum. Par contre, la diversification observée dans le genre Angraecum à Madagascar est le résultat d’une accumulation progressive des lignées et non pas d’une radiation adaptative rapide comme nous l’avions prédit. Finalement, nous avons pu montrer que l’ancêtre d’Angraecum s. s. est malgache et que la dispersion s’est effectuée de Madagascar vers d’autres régions.

Délimitation du genre Angraecum sensu stricto

Notre projet a permis d’accroître l’échantillonnage du genre Angraecum dans l'étude la plus exhaustive du groupe jamais réalisée, incluant près de 70% des espèces malgaches, dont la majorité jamais présentées dans des analyses phylogénétiques auparavant. Cet échantillonnage significatif a permis dans un premier temps de résoudre les problèmes liés à la systématique du groupe, entre autre le statut d’Angraecum comme genre monophylétique. Ceci a permis d’exclure définitivement la plupart des sections africaines qui étaient considérées appartenir au genre Angraecum sensu lato par bon nombre d’auteurs. Deuxièmement, la résolution des clades à l’intérieur d’Angraecum sensu stricto, maintenant monophylétique, a permis de réviser les sections proposées par Garay (1973). Cette révision fut rendue possible grâce à la concordance de nos données moléculaires avec les données morphologiques. La plus grande contribution de notre projet dans l’étude de la systématique du groupe a été la découverte d’un nouveau caractère, la position du labellum, supérieure ou inférieure selon les espèces, qui semble bien délimiter chaque section. Dans cette nouvelle classification, nous reconnaissons 14 sections, dont cinq sont nouvellement décrites: Africanae, Oeoniella, Robusta, Sobennikoffia et Stellariangraecum.

Notre étude a montré l’importance des convergences évolutives au sein de la sous-tribu des Angraecinae, qui jusqu’à maintenant étaient difficiles à observer avec seulement les données morphologiques, ce qui avait rendu difficile l’étude systématique du groupe. La reconnaissance de deux nouveaux genres malgaches, Acaulia et Parangraecum, témoignent de l’existence de caractères homoplasiques dans la sous-tribu. Par ailleurs, malgré la résolution de la phylogénie du genre Angraecum et la quantité de données morphologiques que nous avons généré, nous étions toujours incapables de définir le genre Angraecum sensu stricto correctement. Ceci est surtout dû aux caractères plésiomorphes qui réunissaient différents genres au sein des Angraecinae. D’autres études morphologiques beaucoup plus élargies, incluant plus d’espèces d'Angraecinae, seraient nécessaires afin de délimiter correctement le genre Angraecum et probablement d’autres genres appartenant à la sous-tribu.

Origine et diversité d’Angraecum à Madagascar

Notre étude sur la diversification des Angraecinae appuie l’hypothèse sur la diversification des plantes vasculaires des régions tropicales humides (Couvreur et al., 2011), qui résulterait d’une accumulation d'espèces à travers le temps plutôt que d’une radiation adaptative rapide. Notre étude confirme aussi une hypothèse concernant l’origine de la diversification des orchidées selon laquelle plusieurs caractères et états de caractères seraient responsables de la diversification des espèces (Freudenstein et Chase, 2015.) et non pas un seul trait ou encore moins un seul trait innovateur. Ceci renforce l’importance des données morphologiques dans les analyses macroévolutives. L’analyse intégrale de tous nos caractères morphologiques montre que beaucoup de caractères ont contribué à la diversification d’Angraecum, soit 22 des 39 caractères évalués (Annexe 11). Cette diversification serait en partie liée à l’expansion de la forêt tropicale humide qui a commencé dans la partie Nord Est de Madagascar au moment de l’établissement de la mousson pendant le Miocène. Cette expansion aurait été suivie d'une migration des espèces à l’intérieur de Madagascar, puis plus tard vers l’extérieur. La colonisation à l’intérieure de l’île s’est produite entre le Miocène et le début du Pliocène. Une migration vers les régions des hautes terres (forêts de montagne

103

adjacentes à celles du Nord-Est) s’est produite vers la fin du Miocène, suivie plus tard de migrations indépendantes vers les autres régions de l’île vers le Pliocène-Pléistocène.

Notre étude suggère que Madagascar serait l’origine de la sous-tribu des Angraecinae, ainsi que du genre Angraecum. Ceci confirme bien notre hypothèse de départ. En effet, la présence des espèces d’Angraecum hors de Madagascar est le résultat d’événements de dispersion à longue distance entraînés soit par les vents dominants (alizée et mousson), qui soufflent dans la région ouest de l’Océan Indien, soit par les cyclones qui sévissent chaque année dans cette région. Le modèle de dispersion suit généralement le mouvement et la direction des vents: elle est de direction Est-Ouest pour les vents dominants, et de direction Est-Ouest ou inversement dans le cas des cyclones. L’expansion des espèces dans les îles voisines, entre autres les Comores et les Mascareignes, plus qu’ailleurs, pourrait être liée à l’âge récent de ces îles et aussi à l’affinité de leur composition floristique avec Madagascar.

Perspectives futures

Certes, notre étude aura permis de reconstruire la phylogénie d’Angraecum avec une meilleure résolution au moins au niveau des nœuds (les principaux clades ou sections), mais elle manque encore de résolution au niveau des espèces. Nous suggérons l’utilisation de marqueurs moléculaires autre que chloroplastiques pour avoir plus de support, d'une part, mais aussi pour valider la topologie que nous avons obtenue. A cause des problèmes liés aux champignons endophytes, nous avons eu de la difficulté à amplifier les régions nucléaires, entre autres ITS, pour beaucoup de nos échantillons, raison pour laquelle nous n’avons pas présenté ces résultats. De nouvelles recherches d’amorces plus spécifiques à notre groupe pourraient résoudre le problème en partie. Cependant, Carlsward et al. (2006) avaient rapporté que les séquences paralogues et orthologues sont fréquentes chez les Angraecinae malgaches (e.g. Aeranthes, Angraecum, Jumellea, etc.) pour la région ITS, ce qui nécessiterait beaucoup plus d’attention lorsqu’on fait les reconstructions phylogénétiques.

Étant donnée la diversification du genre Angraecum à Madagascar et dans les îles adjacentes, une étude sur les pollinisateurs pourrait fournir des informations importantes sur

104

les mécanismes de spéciation, d'adaptation et/ou de colonisation de niche. En effet, plusieurs auteurs avancent que l’autopollinisation observée chez certaines espèces d’Angraecum qui se trouvent en dehors de Madagascar (Hermans et Hermans, 2013; Paillet et al., 2013) serait due à l’absence de pollinisateurs spécialisés. Cette absence aurait conduit à une nouvelle adaptation de la plante en vue de maintenir sa reproduction. Nous pensons que le changement de strates écologiques chez Angraecum pourrait être lié à un changement de pollinisateurs. Ceci concerne principalement les espèces à fleurs vertes qui se sont beaucoup diversifiées à Madagascar et dans les Mascareignes (ces espèces constituent plus de la moitié des espèces d’Angraecum dans l’île de la Réunion).

Afin de mieux définir le genre Angraecum, il serait recommandé d'examiner plus de caractères pour renforcer l'analyse des caractères morphologiques et anatomiques des Angraecinae. Des analyses morphométriques pourraient aussi être utile, notamment la mesure de la taille du gynostème ou de la longueur des stipes. Les travaux de Dressler (1996), Freudenstein et Rasmussen (2005), ou autres, sur les caractères morphologiques et anatomiques des orchidées doivent être appuyées pour une meilleure compréhension de l’histoire évolutive de la famille des Orchidaceae.

105

Bibliographie

Acharya, K.P., Vetaas, O.R. & Birks, H.J.B. 2011.Orchid species richness along Himalayan elevational gradients. J. Biogeo. 38: 1821–1833. Akaike, H. 1974. A new look at statistical model identiﬁcation. IEEE Trans. Autom. Control 19: 716–723. Ali, J.R. & Aitchison, J.C. 2009. Kerguelen Plateau and the Late Cretaceous southern- continent bioconnection hypothesis: tales from a topographical ocean. J. Biogeogr. 36: 1778–1784. Ali, J.R. & Huber, M. 2010. Mammalian biodiversity on Madagascar controlled by ocean currents. Nature 463: 653–657. Ali, J.R. & Krause, D.W. 2011. Late Cretaceous bioconnections between Indo-Madagascar and Antarctica: refutation of the Gunnerus Ridge causeway hypothesis. J. Biogeo. 38: 1855–1872. Alsos, I.G., Ehrich, D., Eidesen, P.B., Solstad, H., Westergaard, K.B., Schonswetter, P., Tribsch, A., Birkeland, S., Elven, R. & Brochmann, C. 2015. LDD to North Atlantic islands: colonization routes and founder effect. AoB Plants 7: plv036. DOI:10.1093/aobpla/plv036. Anderson, C.L., Bremer, K. & Friis, E.M. 2005. Dating phylogenetically basal eudicots using rbcL sequences and multiple fossil reference points. Am. J. Bot. 92: 1737–48. Anthony, F., Diniz, L.E.C., Combes, M.C. & Lashermes, P. 2010. Adaptive radiation in Coffea subgenus Coffea L. (Rubiaceae) in Africa and Madagascar. Plant. Syst. Evol. 285: 51–64. DOI: 10.1007/s00606-009-0255-8. APG III. 2009. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG III. Bot. J. Linn. Soc. 161: 105–121. Arditii, J. 1992. Fundamentals of orchid biology. John Wiley and Sons, New York, USA. Arditti, J. & Ghami, A.K.A. 2000. Tanksley Review No. 110: numerical and physical properties of orchid seeds and their biological implications. New Phytol. 145: 367–421. Arditti, J. 2003. Resupination. Lankesteriana 7: 95–96.

Arditii, J. Elliott, J., Kitching, I.J. & Wasserthal, L.T. 2012. Good Heavens what insect can suck it – Charles Darwin, Angraecum sesquipedale and Xanthopan morganii praedicta. Bot. J. Linn. Soc. 169: 403–432. Aslanian, D. & Moulin, M. 2010. Comment on ‘A new scheme for the opening of the South Atlantic Ocean and the dissection of an Aptian salt basin’ by Trond H. Torsvik, Sonia Rousse, Cinthia Labails and Mark A. Smethurst. Geophys. J. Int. 183(1): 20–28. Bacon, C.D., Simmons, M.P., Archer, R.H., Zhao, L.C. & Andriantiana, J. 2016. Biogeography of the Malagasy Celastraceae: Multiple independent origins followed by widespread dispersal of genera from Madagascar. Mol. Phylogenet. Evol. 94: 365–382. DOI: 10.1016/j.ympev.2015.09.013. Baker, J.G. 1877. Orchideae (S. Moore.) Lystrostachys. pp. 354--355 in: Flora of Mauritius and the Seychelles: a description of the flowering plants and ferns of those islands. L. Reeve & Co. London. Baksi, A.K. & Hoffman, K.A. 2000. On the age and morphology of the Reunion event. J. Geophys. Res. 27(18): 2997–3000. Battistini, R. & Richard-Vindard, G. 1972. Biogeography and Ecology in Madagascar. W. Junk B.V., The Hague. Bory de Saint Vincent J.B.G. 1804. Voyage dans les principales iles des mers d’Afrique. F. Buisson, Paris. Bremer, K. 1992. Ancestral areas: a cladistic reinterpretation of the center of origin concept. Syst. Biol. 41: 436–455. Bremer, K. 2000. Early Cretaceous lineages of monocot flowering plants. Proc. Nat. Acad. Sci. USA. 97: 4707–4711. Bremer, K., Friis, E.M. & Bremer, B. 2004. Molecular phylogenetic dating of asterid flowering plants shows Early Cretaceous diversification. Syst. Biol. 53: 469–505. Buerki, S., Devey D.S., Callmander, M.W., Phillipson, P.B. & Forest, F. 2013. Spatio- temporal history of the endemic genera of Madagascar. Bot. J. Linn. Soc. 171: 304– 329. Carlsward, B.S., Whitten, M.W. & Williams, N.H. 2003. Molecular phylogenetics of Neotropical leaﬂess Angraecinae (Orchidaceae): re-evaluation of generic concepts. Int. J. Plant Sci. 164: 43–51.

Carlsward, B.S., Stern, W.L. & Bytebier, B. 2006. Comparative vegetative anatomy and systematics of the angraecoids (Vandeae, Orchidaceae) with an emphasis on the leafless habit. Bot. J. Linn. Soc. 151: 165–218. Chang, Y.Y., Kao, N.H., Li, J.Y., Hsu, W.H. & Liang, Y.L. 2010. Characterization of the possible roles for B-class MADS-box genes in regulation of perianth formation in orchid. J. Plant Physiol. 152: 837–853. Chase, M.W. & Hills, H.G. 1991. Silica gel: an ideal material for field preservation of leaf samples for DNA studies. Taxon 40: 215–220. Chase, M.W., Barrett, R.L., Cameron, K.M., Hills, H.G. & Freudenstein, J.V. 2003. DNA data and Orchidaceae systematics: A new phylogenetic classification. In K. M. Dixon [ed.], Orchid conservation pp. 69--89. Chase, M.W. 2005. Classification of Orchidaceae in the age of DNA data. Curt. Bot. Mag. 22: 2–7. Chase, M.W., Cameron, K.M., Freudenstein, J.V., Pridgeon, A.M., Salazar, G., van den Berg, C. & Schuiteman, A. 2015. An updated classiﬁcation of Orchidaceae. Bot. J. Linn. Soc. 177: 151–174. Chen, S., Yao, H., Han, J., Liu, C., Song, J., Shi, L., Zhu, Y., Ma, X., Gao, T., Pang, X., Luo, K., Li, Y., Li, X., Jia, X., Lin, Y. and Leon, C. 2010. Validation of the ITS2 region as a novel DNA barcode for identifying medicinal plant species. PLoS ONE 5(1): e8613. DOI: 10.1371/journal.pone.0008613. Christidis, L., Goodman, S.M., Naughton, K. & Appleton, B. 2014. Insights into the Evolution of a Cryptic Radiation of Bats: Dispersal and Ecological Radiation of Malagasy Miniopterus (Chiroptera: Miniopteridae). PLoS ONE 9(3): e92440. DOI: 10.1371/journal.pone.0092440. CITES. 2015. Convention on international trade in endangered species of wild fauna and flora, Appendices I and II. http://www.cites.org/eng/app/appendices.php (accessed 5 March 2015). Clark, J.L. & Zimmer, E.A. 2003. A preliminary phylogeny of Alloplectus (Gesneriaceae): implications for the evolution of flower resupination. Syst. Bot. 28(2): 365–375.

iii

Conran, J.G., Bannister, J.M. & Lee, D.E. 2009. Earliest orchid macrofossils: early Miocene Dendrobium andearina (Erchidaceae, Epidendroideae) from New Zealand. Am. J. Bot. 96(2): 466–474. Couvreur, T.L.P., Forest, F. & Baker, W.J. 2011. Origin and global diversification patterns of tropical rain forests: inferences from a complete genus-level phylogeny of palms. BMC Biol. 9: 44. http://www.biomedcentral.com/1741-7007/9/44 Cribb, P.J., Hermans, J. & Roberts, D.L. 2007. Erasanthe (Orchidaceae. Epidendroideae, Vandeae, Aerangidinae), a new endemic orchid genus from Madagascar. Adansonia 3(29): 27–30. Cribb, P.J. & Hermans, J. 2009. Field guide to the orchids of Madagascar. Kew, Kew Publishing. Cribb, P.J. & Carlsward, B. 2012. New combinations in Aerangis, Diaphananthe and Podangis (Orchidaceae, subtribe Angraecinae). Phytotaxa 71: 42–47. DOI: 10.11646/phytotaxa.71.1.8. Crisci, J.V. 2001. The voice of historical biogeography. J. Biogeogr. 28: 157–158. Davis, C.C., Webb, C.O., Wurdack, K.J., Jaramillo, C.A. & Donoghue, M.J. 2005. Explosive radiation of Malpighiales supports a mid-Cretaceous origin of modern tropical rain forests. Am. Nat. 165: e36–65. Ding L., Wang, Y. & Yu, H. 2013. Overexpression of DOSOC1, an ortholog of Arabidopsis SOC1, promotes flowering in the orchid Dendrobium Chao Parya Smile. Plant Cell Physiol. 54(4): 595–608. Donque G., 1975. Contribution géographique à l'étude du climat de Madagascar. Nouv. Imp. Art. Graph. Tananarive, 478 pages. Doyle, J.J. & Doyle, J.L. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochem. Bull. 19: 11–15. Dressler, R.L. 1981. The Orchids, Natural History and Classification. Harvard University Press, Cambridge. Dressler, R.L. 1993. Phylogeny and Classiﬁcation of the Orchid Family. Dioscorides Press, Portland.

Drummond, A.J. & Rambaut, A. 2007. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol. Biol. 7: 214. Du Puy, D.J. & Moat J. 1996. A refined classification of the primary vegetation of Madagascar based on the underlying geology: using GIS to map its distribution and to assess its conservation status. Biogéographie de Madagascar pp. 205--218. Du Puy, D., Cribb, P.J., Bosser, J., Hermans, J. & Hermans, C. 1999. The Orchids of Madagascar. First edition. Kew, Royal Botanical Gardens. Eagles, G. & Konig, M. 2008. A model of plate kinematics in Gondwana breakup. Geophy. J. Internat. 173: 703–717. DOI: 10.1111/j.1365-246X.2008.03753. Elhaih. 1968. Biogéographie. Paris, Armand Colin, coll. U. 406 pages. Faramalala, M.H. 1995. Formations végétales et domaine forestier national de Madagascar. Conservation Internationa (et al.), 1 map. Fischer, G.A., Gravendeel, B., Sieder A., Andriantiana, J., Heiselmayer, P., Cribb, P.J., Smidt, E.C., Samuel, R. & Kiehn, M. 2007. Evolution of resupination in Malagasy species of Bulbophyllum (Orchidaceae). Mol. Phylogenet. Evol. 45: 358–376. FitzJohn, R.G., Maddison, W.P. & Otto, S.P. 2009. Estimating trait-dependent speciation and extinction rates from incompletely resolved phylogenies. Syst. Biol. 58: 595–611. FitzJohn, R.G. 2010. Quantitative traits and diversiﬁcation. Syst. Biol. 59: 619–633. FitzJohn, R.G. 2012. Diversitree: comparative phylogenetic analyses of diversiﬁcation in R. Meth. Eco. Evol. 3: 1084–1092. Freudenstein, J.V. & Rasmussen, F.N. 1999. What does morphology tell us about orchid relationships? A cladistic analysis. Am. J. Bot. 86(2): 225–248. Freudenstein, J.V. & Chase, M.W. 2015. Phylogenetic relationships in Epidendroideae (Orchidaceae), one of the great flowering plant radiations: progressive specialization and diversification. Annal. Bot. 115: 665–681. DOI:10.1093/aob/mcu253. Galochet, M. 2008. Histoire de la biogéographie française des origines à nos jours. 18 pages. Garay, L.A. 1973. Systematics of the genus Angraecum (Orchidaceae). Kew Bull. 28: 495– 516.

Gavrilets, S. & Vose, A. 2005. Dynamic patterns of adaptive radiation. Proc. Natl. Acad. Sci. USA 102: 18040–18045. Givnish, T.J., Spalink, D., Ames, M., Lyon, S.P., Hunter, S.J., Zuluaga, A., Iles, W.J.D., Clements, M.A., Arroyo, M.T.K., Leebens-Mack, J., Endara, L., Kriebel, R., Neubig, K.M., Whitten, W.M., Williams, N.H. & Cameron, K.M. 2015. Orchid phylogenomics and multiple drivers of their extraordinary diversification. Proc. R. Soc. B. 282: 20151553. DOI: 10.1098/rspb.2015.1553. Goldberg, J., Knapp, M., Emberson, R.M., Townsend, J.I. & Trewick, S.A. 2014. Species radiation of carabid beetles (Broscini: Mecodema) in New Zealand. PLoS ONE 9(1): e86185. DOI: 10.1371/journal.pone.0086185. Goodman, S.M. & Benstead, J.P. 2005. Updated estimates of biotic diversity and endemism for Madagascar. Oryx 39: 73–77. Goodman, K.R., Kelley, J.P., Welter, S.C., Roderick, G.K. & Elias, D.O. 2015. Rapid diversification of sexual signals in Hawaiian Nesosydne planthoppers (Hemiptera: Delphacidae): the relative role of neutral and selective forces. J. Evol. Biol. 28: 415– 427. Gouy, M., Guindon, S. & Gascuel, O. 2010. SeaView version 4: a multiplatform graphical user interface for sequence alignment and phylogenetic tree building. Mol. Biol. Evol. 27(2): 221–224. Govaerts, R., Campacci, M.A., Baptista, D.H., Cribb, P.J., George, A., Kreuz, K. & Wood, J. 2015. World checklist of Orchidaceae. Facilitated by the Royal Botanic Gardens, Kew. Available at: http://apps.kew.org/wcsp/ (accessed 22 March 2015). Grubb, P.J. 2003. Interpreting some outstanding features of the flora and vegetation of Madagascar. Perspect. Plant Ecol. Evol. Syst. 6: 125–46. Gustafsson, A.L.S., Verola, C.F. & Antonelli, A. 2010. Reassessing the temporal evolution of orchids with new fossils and a Bayesian relaxed clock, with implications for the diversification of the rare South American genus Hoffmannseggella (Orchidaceae: Epidendroideae). BMC Evol. Biol. 10: 177.

Hall, T.A. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symp. Ser. 41: 95–98. Heibl, C. 2008 (last updated 04 September 2013). PHYLOCH: R language tree plotting tools and interfaces to diverse phylogenetic software packages. http://www.christophheibl.de/Rpackages.html (accessed 22 August 2015). Hermans, J. & Cribb, P. 2005. The Angraecum germinyanum complex. Orchid Rev. 113: 90– 97. Hermans, J., Hermans, C., Du Puy, D., Cribb, P. & Bosser J. 2007. Orchids of Madagascar. 2nd edition. R. Bot. Gard. Kew. Hermans, C. & Hermans, J. 2013. In search of the star of Africa. Orchids Bull. AOS 82(9): 544–550. Hossain, I. & Levy, C. 2014. Investigating KNOX gene expression in Aquilegia petal spur development. J. Emerg. Investig. 1–6. Hsu, H.F., Huang, C.H., Chou, L.T. & Yang, C.H. 2003. Ectopic expression of an orchid (Oncidium Gower Ramsey) agl6-like gene promotes ﬂowering by activating ﬂowering time genes in Arabidopsis thaliana. Plant Cell Physiol. 44: 783–794. Huelsenbeck, J.P. & Ronquist, F. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinfo. 19: 1572–1574. Humbert, H. 1955. Les Territoires Phytogéographiques de Madagascar. Leur artographie. Colloque sur les Régions Écologiques du Globe, Paris 1954. Ann. Biol., 31: 195–204, + map. Iles, W.J.D., Smith, S.Y., Gandolfo, M.A. & Graham, S.W., 2015. Monocot fossils suitable for molecular dating analyses. Bot. J. Linn. Soc. 178(3): 346–374. IUCN. 2015. The International Union for Conservation of Nature red list of threatened species. http://www.iucnredlist.org. Jacquemyn, H., Micheneau, C., Roberts, D.L. & Pailler, T. 2005. Elevational gradients of species diversity, breeding system and floral traits of orchid species on Reunion Island. J. Biogeogr. 32: 1751–1761. DOI:10.1111/j.1365-2699.2005.01307. Janssen, T. & Bremer, B. 2004. The age of major monocot groups inferred from 800+ rbcL sequences. Bot. J. Linn. Soc. 146: 385–398.

vii

Janssen, T., Bystriakova, N., Rakotondrainibe, F., Coomes, D., Labat, J.N. & Schneider, H., 2008. Neoendemism in Madagascan scaly tree ferns results from recent, coincident diversification bursts. Evolution 62(8): 1876–1889. Jermy, T. 1999. Deep flowers for long tongues: a final word. Trends Ecol. Evol. 14: 34. Jersakova, J. & Malinova, T. 2007. Spatial aspects of seed dispersal and seedling recruitment in orchids. New Phytologist 176: 235–237. Johnson, S.D. & Steinerb, K.E. 2000. Generalization versus specialization in plant pollination systems. Trends Ecol. Evol. 15(4): 140–143. DOI: 10.1016/S0169-5347(99)01811. Joly, F. 1941. Les climats de Madagascar. In: L'information géographique 5(4): 76–80. DOI: 10.3406/ingeo.1941.5092. Jonsson, K.A., Fabre, P.H., Fritz, S.A., Etienne, R.S., Ricklefs, R.E., Jorgensen, T.B., Fjeldsa, J., Rahbek, C., Ericson, P.G.P., Woog, F., Pasquet, E. & Irestedt, M. 2012. Ecological and evolutionary determinants for the adaptive radiation of the Madagascan vangas. PNAS 109(17): 6620–6625. Kembel, S.W., Ackerly, D.D., Blomberg, S.P., Cornwell, W.K., Cowan, P.D., Helmus, M.R., Morlon, H. & Webb, C.O. 2015. Package ‘picante’: R tools for integrating phylogenies and ecology, Version 1.6-2. https://cran.r-project.org/web/packages/picante/index.html (accessed 15 August 2015). Kerbauy, G.B., Takahashi, C.A., Lopez, A.M., Matsumura, A.T., Hamachi, L., Félix, L.M., Pereira, P.N., Freschi, L. & Mercier, H. 2012. Crassulacean Acid Metabolism in Epiphytic Orchids: Current Knowledge, Future Perspectives. Applied photosynthesis pp. 81--104. Kluge, M., Vinson, B. & Ziegler, H. 1997. Ecophysiological studies on orchids of Madagascar: incidence and plasticity of crassulacean acid metabolism in species of the genus Angraecum Bory. Plant. Ecol. 135: 43–57. Krüger, A., Razafimandimbison, S.G. & Bremer, B. 2012. Molecular phylogeny of the tribe Danaideae (Rubiaceae: Rubioideae): Another example of out-of-Madagascar dispersal. Taxon 61(3): 629–636. Kuntner, M. & Agnarsson, I. 2011. Phylogeography of a successful aerial disperser: the golden orb spider Nephila on Indian Ocean islands. BMC, Evol. Biol. 11: 119.

viii

Leroy, J.F. 1978. Composition, origin, and affinities of the Madagascan vascular flora. Ann. Mo. Bot. Gard. 65: 535–89. Lewis, D.S., Sperling, F.A.H., Nakahara, S., Cotton, A.M., Kawahara, A.Y. & Condamine, F.L. 2015. Role of Caribbean Islands in the diversification and biogeography of Neotropical Heraclides Swallowtails. Cladistics 31: 291–314. Luo, Y., Widmer, A. & Karrenberg, S. 2015. The roles of genetic drift and natural selection in quantitative trait divergence along an altitudinal gradient in Arabidopsis thaliana. Heredity 114: 220–228. MacArthur, R.H. & Wilson, E.O. 1967. The Theory of Island Biogeography. Princeton University Press, Princeton, NJ. Maddison, W.P., Midford, P.E. & Otto, S.P. 2007. Estimating a binary character’s eﬀect on speciation and extinction. Syst. Biol. 56: 701–710. Maddison, W.P. & FitzJohn, R.G. 2015. The unsolved challenge to phylogenetic correlation tests for categorical characters. Syst. Biol. 64: 127–136. Mallet, B., Martos, F., Blambert, L., Pailler, T. & Humeau, L. 2014. Evidence for Isolation- by-Habitat among Populations of an Epiphytic Orchid Species on a Small Oceanic Island. PLoS ONE 9(2): e87469. DOI: 10.1371/journal.pone.0087469. Martin, R.D. 1972. Adaptive radiation and behaviour of the Malagasy lemurs. Philos. Trans. R. Soc. Lond. B Biol. Sci. 264: 295–352. Matzke, N.J. 2013a. Probabilistic historical biogeography: new models for founder-event speciation, imperfect detection, and fossils allow improved accuracy and model- testing. Thesis, University of California, Berkeley, 251 pages. Matzke, N.J. 2013b. BioGeoBEARS: BioGeography with Bayesian (and Likelihood) Evolutionary Analysis in R Scripts. Berkeley, CA: University of California. http://CRAN.R-project.org/package=BioGeoBEARS. Matzke, N. J. 2014. Model selection in historical biogeography reveals that founder-event speciation is a crucial process in island clades. Syst. Biol. 63: 951–970. Micheneau, C. 2005. Systématique moléculaire de la sous-tribu des Angraecinae (Vandae, Orchidaceae): Perspective taxonomique et implications de la relation plante-

pollinisateurs dans l’évolution des formes florales réunionnaises. Thèse de doctorat, Université de la Réunion, 360 pages. Micheneau, C., Fournel, J. & Pailler, T. 2006. Bird pollination in an Angraecoid orchid on Reunion Island (Mascarene Archipelago, Indian Ocean). Ann. Bot. (Lond.) 97: 965– 974. Micheneau, C., Carlsward, B., Fay, M.S., Bytebier, B., Pailler, T. & Chase, M.W. 2008a. Phylogenetics and biogeography of Mascarene angraecoid orchids (Vandeae, Orchidaceae). Mol. Phylogenet. Evol. 46: 908–922. Micheneau, C., Fournel, J., Gauvin-Bialecki, A. & Pailler, T. 2008b. Auto-pollination in a long-spurred endemic orchid (Jumellea stenophylla) on Reunion Island (Mascarene Archipelago, Indian Ocean). Plant Syst. Evol. 272: 11–22. Micheneau, C., Fournel, J., Humeau, L. & Pailler, T., 2008c. Orchid–bird interactions: a case study from Angraecum (Vandeae, Angraecinae) and Zosterops (white-eyes, Zosteropidae) on Reunion Island. Botany 86: 1143–1151. Micheneau, C., Johnson, S.D. & Fay, M.F. 2009. Orchid pollination: from Darwin to the present day. Bot. J. Linn. Soc. 161: 1–19. Micheneau, C., Fournel, J., Warren, B.H., Hugel, S., Gauvin-Bialecki, A., Pailler, T., Strasberg, D. & Chase, M.W. 2010. Orthoptera, a new order of pollinator. Ann. Bot. (Lond.) 105: 355–364. Moen, D. & Morlon, H. 2012. Why does diversification slow down? Trends Ecol. Evol. 29(4): 190–197. Moya, A., Galliana, A. & Ayala, F.J. 1995. Founder-effect speciation theory: failure of experimental corroboration. Proc. Natl. Acad. Sci. USA. 92: 3983–3986. Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B. & Kent, J. 2000. Biodiversity hotspots for conservation priorities. Nature 403: 853–858. Nilsson, L.A., Jonsson, L., Rason, L. & Randrianjohany, E. 1985. Monophily and pollination mechanisms in Angraecum arachnites Schltr. (Orchidaceae) in a guild of long-tongued hawk-moths (Sphingidae) in Madagascar. Biol. J. Linn. Soc. 91: 1–19. Nilsson, L.A., Jonsson, L., Ralison, L. & Randrianjohany, E. 1987. Angraecoid orchids and hawkmoths in central Madagascar: specialized pollination systems and generalist foragers. Biotropica 19: 310–318.

Nilsson, L.A. 1998. Deep flowers for long tongues: reply from L. A. Nilsson. Trends Ecol. Evol. 13: 509. Otero, J.T. & Flanagan, N.S. 2006. Orchid diversity beyond deception. Trends Ecol. Evol. 21: 64–65. Oxelman, B., Lidén, M. & Berglund, D. 1997. Chloroplast rps16 intron phylogeny of the tribe Sileneae (Caryophyllaceae). Plant Syst. Evol. 206: 393–410. Pailler, T., Tournebize, R. & Henze, F. 2013. Guide des orchidées de la Réunion, 52 espèces indigènes décrites. 5ème colloque internationale sur la conservation des orchidées. 1ère ed. Université de la Réunion. Pagel, M. 1994. Detecting correlated evolution on phylogenies: a general method for the comparative analysis of discrete characters. Proc. R. Soc. Ser B. 255: 37–45. Pan, Z.J., Cheng, C.C., Tsai, W.C., Chung, M.C., Chen, W.H., Hu, J.M. & Chen, H.H. 2011. The duplicated B-class MADS-box genes display dualistic characters in orchid floral organ identity and growth. Plant Cell Physiol. 52(9): 1515–1531. Perrier de la Bâthie, H. 1941. Flore de Madagascar (plantes vasculaires): 49ème Famille– Orchidées. Tome II ed. H. Humbert. Imprimerie Officielle, Tananarive pp. 202--321. Peter, C.I. & Johnson, S.D. 2014. A pollinator shift explains ﬂoral divergence in an orchid species complex in South Africa. Ann. Bot. (Lond.) 113: 277–288. Phillimore, A.B. & Price, T.D., 2008. Density-dependent cladogenesis in birds. PLoS ONE 6: e71. DOI: 10.1371/journal.pbio.0060071. Porembski, S. & Barthlott, W. 1988. Velamen radicum micromorphology and classification of Orchidaceae. Nordic J. Bot. 8: 117–137. Posada, D. & Crandall, K. 1998. Modeltest: testing the model of DNA substitution. Bioinformatics (Oxf.) 14: 817–818. Pybus, O.G. & Harvey, P.H. 2000. Testing macro-evolutionary models using incomplete molecular phylogenies. Proc. Biol. Sci. 267: 2267–2272. de Queiroz, A. 2005. The resurrection of oceanic dispersal in historical biogeography. Trends Ecol. Evol. 20, 68–73. Rabinowitz, P.D. & Woods S. 2006. The Africa–Madagascar connection and mammalian migrations. J. Afric. Earth Sci. 44: 270–276.

Rabosky, D.L., Santini, F., Eastman, J.T., Smith, S.A., Sidlauskas, B., Chang, J. & Alfaro, M.E. 2013. Rates of speciation and morphological evolution are correlated across the largest vertebrate radiation. Nat. Commun. 4: 1958. Rabosky, D.L. 2014. Automatic detection of key innovations, rate shifts, and diversity dependence on phylogenetic trees. PLoS ONE 9: e89543. DOI: 10.1371/journal.pone.0089543. Rabosky, D.L., Donnellan, S.C., Grundler, M. & Lovette, I.J. 2014. Analysis and visualization of complex macroevolutionary dynamics: an example from Australian scincid lizards. Syst. Biol. 63(4): 610–627. Rabosky, D.L. & Goldberg, E.E. 2015. Model inadequacy and mistaken inferences of trait- dependent speciation. Syst. Biol. 64(2): 340–355. Rakotoarinivo, M., Blach-Overgaard, A., Baker, W.J., Dransfield, J., Moat, J. & Svenning, J.C. 2013. Palaeoprecipitation is a major determinant of palm species richness patterns across Madagascar: a tropical biodiversity hotspot. Proc. R. Soc. B 280: 1–9. DOI: 10.1098/rspb.2012.3048. Rakotoarivelo, F.P., Razafimandimbison, S.G., Mallet, B., Faliniaina, L. & Pailler T. 2012. Molecular systematics and evolutionary trends and relationships in the genus Jumellea (Orchidaceae): Implications for its species limits. Taxon 61(3): 534–544. Ramirez, S.R., Gravendeel, B., Singer, R.B., Marshall, C.R. & Pierce, N.E. 2007. Dating the origin of the Orchidaceae from a fossil orchid with its pollinator. Nature 448: 1042– 1045. Raxworthy, C.J. & Nussbaum, R.A. 1995. Systematics, speciation and biogeography of the dwarf chameleons (Brookesia, Reptilia, Squamata, Chamaeleontidae) of northern Madagascar. J. Zool. 235: 525–558. DOI:10.1111/j.1469-7998.1995.tb01767. Raxworthy, C.J., Forstner, M.R.J. & Nussbaum, R.A. 2002. Chameleon radiation by oceanic dispersal. Nature 415: 784–787. Reddy, S., Driskell, A., Rabosky, D.L., Hackett, S.J. & Schulenberg, T.S. 2012. Diversification and the adaptive radiation of the vangas of Madagascar. Proc. R. Soc. B. 279: 2062–2071.

xii

Ree, R.H. & Smith, S. A. 2008. Maximum likelihood inference of geographic range evolution by dispersal, local extinction, and cladogenesis. Syst. Biol. 57: 4–14. Richard, A. 1828. Monographie des orchidées des iles de France et Bourbon. Mémoire de la société d’Histoire Naturelle de Paris, Extrait 4. RHS. 2012. Orchid hybrid list. Royal horticultural society. http://www.rhs.org.uk/Plants/RHS- Publications/Orchid-hybrid-lists (accessed 5 March 2015) Ronquist, F. 1997. Dispersal-vicariance analysis: a new approach to the quantification of historical biogeography. Syst. Biol. 46: 195–203. Rudall, P. & Bateman, R.M. 2002. Roles of synorganisation, zygomorphy and heterotopy in foral evolution: the gynostemium and labellum of orchids and other lilioid monocots. Biol. Rev. 77: 403–441. Ruddiman, W.F. 2001. Earth’s climate: past and future, 2nd ed. New York: W. H. Freeman and Company. Scantlebury, D.P. 2013. Diversification rates have declined in the Malagasy herpetofauna. Proc. R. Soc. B. 280: 20131109. Schatz, G.E. 1996. Malagasy/Indo-australo-malesian phytogeographic connections. Biogéographie de Madagascar pp. 73--83. Schlechter, R. 1918. Versuch einer naturlichen Neuordnung der afrikanischen agraekoiden Orchidaceen. Beih. Bot. Centralbl. 36: 62–181. Schlechter, R. 1925. Orchidaceae Perrierianae. Fedde Rep. Beih. 33: 1–391. Schluter, D. 2000. The ecology of adaptive radiation. Oxford University Press Inc., New York. Shi, J.J. & Rabosky, D.L. 2015. Speciation dynamics during the global radiation of extant bats. Evolution 69(6): 1528–1545. Silva, W.A.Jr., Costa, M.C., Valente, V., Sousa, J.F., Pacó-Larson, M.L., Espreafico, E.M., Camargo, S.S., Monteiro, E., Holanda, A.J., Zago, M.A., Simpson, A.J. & Dias Neto, E. 2001. PCR template preparation for capillary DNA sequencing. BioTechniques 30: 537–542. Simo-Droissart, M., Micheneau, C., Sonke, B., Droissart, V., Plunkett, G.M., Lowry, P.P., Hardy, O.J. & Stevart, T. 2013. Morphometrics and molecular phylogenetics of the continental-African species of Angraecum section Pectinaria (Orchidaceae). Plant Ecol. Evol. 146: 295–309.

xiii

Sin, H.C., Wing, Y.T. & Arditii, J. 2002. Biology of Vanda Miss Joaquim. Singapore University press, Singapore. Stankiewicz, J., Thiart, C., Masters, J.C. & de Wit, M.J. 2006. Did lemurs have sweepstake tickets? An exploration of Simpson’s model for the colonization of Madagascar by mammals. J. Biogeogr. 33: 221–235. Stévart, T., Cawoy, V., Damen, T. & Droissart, V. 2010. Taxonomy of Atlantic Central African Orchids 1. A new species of Angraecum sect. Pectinaria (Orchidaceae) from Gabon and Equatorial Guinea. Syst. Bot. 35(2): 252–256. Stewart, J., Hermans, J. & Campbel, B. 2006. Angraecoid orchids. Species from the African region. Timber press Inc., Portland. Stoddart, D.R. 1984. Biogeography and ecology of the Seychelles Islands. W. Junk Publ., The Hague, Boston, Lancaster. Strijk, J.S., Noyes, R.D., Strasberg, D., Cruaud, C., Gavory, F., Chase, M.W., Abbott, R.J. & Thebaud, C. 2012. In and out of Madagascar: dispersal to peripheral islands, insular speciation and diversification of Indian Ocean daisy trees (Psiadia, Asteraceae). PLoS ONE 7: e42932. Summerhayes, V.S. 1958. African orchids, XXVI. Kew Bull. 13: 260–277. Swofford, D.L. 2002. PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods), version 4.0 Beta. Sunderland, Massachusetts: Sinauer. Szlachetko, D.L. & Romowicz, A. 2007. Dolabrifolia, un nouveau genre d’orchidées de l’alliance Angraecum. Richardiana 7(2): 53–54. Szlachetko, D.L., Tukałło, P., Mytnik-Ejsmont, J. & Grochocka, E. 2013. Reclassification of the Angraecum alliance (Orchidaceae, Vandoideae) based on molecular and morphological data. Biodiv. Res. Conserv. 29: 1–23. Taberlet, P., Gielly, L., Pautou, G. & Bouvet, J. 1991. Universal primers for ampliﬁcation of three non-coding regions of chloroplast DNA. Plant Mol. Biol. 17: 1105–1109. Thébaud, C., Warren, B.H., Cheke, A. & Strasberg, D. 2009. Mascarene Islands, biology. In: Gillespie, R.G. & Clague, D.A., editor. Encyclopedia of Islands: California University Press, Berkeley pp. 612--619. Theißen G. & Saedler, H. 1995. MADS-box genes in plant ontogeny and phylogeny: Haeckel’s ‘biogenetic law’ revisited. Curr. Opin. Genet. Dev. 5: 628–639.

xiv

DOI: 10.1016/0959-437X(95)80032-8. Torsvik, T.H., Rousse, S. & Smethurst, M.A. 2010. Reply to comment by D. Aslanian and M. Moulin on ‘A new scheme for the opening of the South Atlantic Ocean and the dissection of an Aptian salt basin’. Geophys. J. Int. 183, 29–34. Townsend, T.M., Vieites, D.R., Glaw, F. & Vences, M. 2009. Testing species-level diversification hypotheses in Madagascar: the case of microendemic Brookesia leaf chameleons. Syst. Biol. 58(6): 641–656. Vences, M., Freyhof, J., Sonnenberg, R., Kosuch, J. &Veith, M. 2001.Reconciling fossils and molecules: Cenozoic divergence of cichlid fishes and thebiogeography of Madagascar. J. Biogeogr. 28: 1091–1099. Vences, M., Wollenberg, K.C., Vieites, D.R. & Lees, D.C., 2009. Madagascar as a model region of species diversification. Trends Ecol. Evol. 24: 456–465. Warren, B.H., Strasberg, D., Bruggemann, J.H., Prys-Jones, R.P. & Thébaud, C. 2010.Why does the biota of the Madagascar region have such a strong Asiatic flavour? Cladistics 26: 526–538. DOI: 10.1111/j.1096-0031.2009.00300. Wasserthal, L.T. 1993. Swing-hovering combined with long tongue in hawkmoths, an antipredator adaptation during flower visits. In: Barthlott W, Naumann CM, Schnidt- Loske K, Schuchmann KL, eds. Animal–plant interaction in tropical environments. Annual Meeting of the German Society for Tropical Ecology. Bonn: Zoolo. Forsch. Museum Alex. Koenig pp. 77--87. Wasserthal, L.T. 1997. The pollinators of the Malagasy star orchids Angraecum sesquipedale, A. sororium and A. compactum and the evolution of extremely long spurs by pollinator shift. Bot. Act. 110: 343–359. Weyeneth, N., Goodman, S.M., Appleton, B., Wood, R. & Ruedi, M. 2011. Wings or winds: inferring bat migration in a stepping-stone archipelago. J. Evol. Biol. 24: 1298–1306. DOI: 10.1111/j.1420-9101.2011.02262. Weyeneth, N., Goodman, S.M., Stanley, W.T. & Ruedi M. 2008. The biogeography of Miniopterus bats (Chiroptera: Miniopteridae) from the Comoro Archipelago inferred from mitochondrial DNA. Mol. Ecol. 17: 5205–5219.

Whitman, M., Medler, M., Randriamanindry, J.J. & Rabakonandrianina, E. 2011. Conservation of Madagascar’s granite outcrop orchids: the influence of fire and moisture. Lankesteriana 11(1): 55–67. Whittaker, R.J., Triantis, K.A. & Ladle, R.J. 2008. A general dynamic theory ofoceanic island biogeography. J. Biogeogr. 35: 977–994. Wikström, N., Savolainen, V. & Chase, M.W. 2001. Evolution of the angiosperms: calibrating the family tree. Proc. R. Soc. B 268: 2211–2220. DOI: 10.1098/rspb.2001.1782. Yoder, A.D., Cartmill, M., Ruvolo, M., Smith, K. & Vilgalyst, R. 1996. Ancient single origin for Malagasy primates. Evolution (93): 5122–5126. Yoder, A.D. & Nowak, M. 2006. Has vicariance or dispersal been the predominant biogeographic force in Madagascar? Only time will tell. Annu. Rev. Ecol. Evol. Syst. 37: 405–31. DOI: 10.1146/annurev.ecolsys.37.091305.110239. Yu, J., Xue, J.H. & Zhou, S.L. 2011. New universal matK primers for DNA barcoding angiosperms. J. Syst. Evol. 49 (3): 176–181.

xvi

Annexe 1. Voucher and sources of DNA data

Taxa, geographical origin, voucher information, and NCBI or BOLD accession numbers (matK, rps16, trnL, ITS) for samples used in the study; sequences that are from different individuals are in italics with voucher information in brackets. Dashes (-) denote unavailable sequences, “unvouchered” are sequences from living collections used in previous studies (Carlsward et al., 2006; Micheneau et al., 2008a), asterisks (*) indicates newly generated sequences, and dagger (†) those with vouchers. Unvouchered specimens have been identified by H.N. Andriananjamanantsoa and J. Andriatiana (TAN). Abbreviations: Com, Comoros; Md, Madagascar; Reu, Reunion. Acampe ochracea (Lindl.) Hochr., Asia, Carlsward 206 (SEL), DQ091314, –, DQ091438, DQ091707; Aerides odorata Lour., Asia, Chase 15081 (K), KF557954, KF558031, KF558231, –; Aerangis ellisii (B.S.Williams) Schltr., Md, Chase 15080 (K), KF557992, KF558043, KF558204, DQ091602 (Carlsward 401, FLAS); Aerangis hariotiana (Kraenzl.) P.J.Cribb & Carlsward, Com, Carlsward 227 (FLAS), DQ091343, –, DQ091467, DQ091606; Aerangis hildebrandtii (Rchb.f.) P.J.Cribb & Carlsward, Com, Kew 2616 (K), DQ091344, –, DQ091468, DQ091608; Aerangis macrocentra (Schltr.) Schltr., Md, Hermans 779 (K), KF557923, –, KF558085, DQ091601 (Kew 779, K); Aerangis punctata J.Stewart, Md, unvouchered (REU), KF557936, KF558040, KF558209, DQ091605; Aeranthes aemula Schltr., Md, Andriananjamanantsoa & al. 00203471 (MT), KT826826*, KT826895*, KT826964*, –; Aeranthes arachnites (Thouars) Lindl., Md, Fournel 126 (REU), KF557945, KF558036, KF558232, DQ091759 (Carlsward 198, FLAS); Aeranthes caudate Rolfe, Md, Chase 17918 (K), KF557929, KF558090, KF558206, –; Aeranthes grandiflora Lindl., Md, unvouchered (K), KF557958, KF558050, KF558158, DQ091760 (Carlsward 238, FLAS); Aeranthes longipes Schltr., Md, DuPuy 17 (K), KF557965, KF558094, KF558208, –; Aeranthes moratii Bosser, Md, Chase 14643 (K), KF558024, KF558060, KF558225, –; Aeranthes neoperrieri Toill.Gen., Md, unvouchered (K), KF557953, KF558059, KF558199, –; Aeranthes strangulata Frapp. ex Cordem., Reu, Pailler 100 (REU), KF558005, KF558097, KF558183, –; Aeranthes tenella Bosser, Reu, Pailler 137 (REU), KF557997, KF558117, KF558162, –; Aeranthes virginalis D.L.Roberts, Com, Chase 17901 (REU), KF557955, KF558071, KF558166, –; Ancistrorhynchus capitatus (Lindl.) Summerh., Africa, Carlsward 276 (FLAS), DQ091351, –, DQ091475, DQ091643; Ancistrorhynchus recurvus Finet, Africa, unvouchered, DQ091354, –, DQ091478, DQ091646; Angraecoides chevalieri (Summerh.) Szlach., Africa, Carlsward 208 (FLAS), AF506363, –,

AF506339, AF506320; Angraecoides cultriformis (Summerh.) Szlach., Africa, Carlsward 298 (FLAS) AF506364, –, AF506340, AF506321; Angraecoides erecta (Summerh.) Szlach., Africa, Bytebier 801 (EA), DQ091323, –, DQ091447, DQ091566; Angraecoides cf. moandense De Wild., Africa, unvouchered (K), KF557944, KF558114, KF558214, –; Angraecopsis parviflora (Thouars) Schltr., Africa, Hermans 4363 (Personal), KF557967, KF558046, KF558150, –; Angraecum amplexicaule Toill.Gen. & Bosser, Md, Andriananjamanantsoa & al. 00203468 (MT), KT826791*, KT826861*, KT826929*, –; Angraecum ankeranense H.Perrier, Md, Andriananjamanantsoa & al. 00203430 (MT), KT826802*, KT826872*, KT826940*, –; Angraecum appendiculatum Frapp. ex Cordem., Reu, Pailler 107 (REU), KF557933, KF558130, KF558173, DQ091752 (Kew 4232, K); Angraecum arachnites Schltr., Md, Hermans 4241 (K), KF557990, KF558064, KF558250, –; Angraecum baronii† (Finet) Schltr., Md, Andriananjamanantsoa & al. 00203336 (MT), KT826787*, KT826857*, KT826925*, –; Angraecum borbonicum Bosser, Md, Andriananjamanantsoa & al. 00203441 (MT), KT826808*, KT826878*, KT826946*, –; Angraecum bracteosum Balf.f. & S.Moore, Reu, Micheneau 7 (REU), KF557998, KF558045, KF558266, –; Angraecum cadetii Bosser, Reu, Pailler 157 (REU), KF557986, KF558122, KF558222; Angraecum calceolus Thouars, Md, Andriananjamanantsoa & al. 00203470 (MT), KT826789*, KT826859*, KT826927*, –; Angraecum caricifolium† H.Perrier, Md, Andriananjamanantsoa & al. 00203338 (MT), KT826776*, KT826846*, KT826914*, –; Angraecum caulescens Thouars, Md, Pailler 97 (REU), KF557988, –, KF558172, –; Angraecum cf. breve† Schltr., Md, Andriananjamanantsoa & al. 00203464 (MT), KT826803*, KT826873*, KT826941*, –; Angraecum cf. peyrotii Bosser, Md, Andriananjamanantsoa & al. 00203426 (MT), KT826806*, KT826876*, KT826944*, –; Angraecum cf. elephantinum† Schltr., Md, Andriananjamanantsoa & al. 00203323 (MT), KT826811*, KT826881*, KT826949*, –; Angraecum cf. germinyanum Hook.f., Hermans 5540 (K), KF557973, KF558093, KF558203, –; Angraecum cf. humile† Summerh., Md, Andriananjamanantsoa & al. 00203453 (MT), KT826762*, KT826833*, KT826902*, –; Angraecum cf. pauciramosum Schltr., Md, Andriananjamanantsoa & al. 00203436 (MT), KT826777*, KT826847*, KT826915*, –; Angraecum cf. humblotianum Schltr., Md, Simo-Droissart & al. 2151 (BRLU), KF672272, KF672255, KF662345, –; Angraecum cf. liliodorum Frapp. ex Cordem., Md, Andriananjamanantsoa & al. 00203472 (MT), KT826801*, KT826871*, KT826939*, –; Angraecum cf. panicifolium H.Perrier, Md, Andriananjamanantsoa & al. 00203421 (MT), KT826816*, KT826885*, KT826954*, –; Angraecum cf. pinifolium† Bosser, Md, Andriananjamanantsoa & al. 00203456 (MT), KT826784*, KT826854*, KT826922*, –; Angraecum cf. rutenbergianum Kraenzl., Md, unvouchered (K), KF557938, KF558076, KF558218, DQ091743 (Carlsward 300, FLAS);

Angraecum cf. sacculatum† Schltr., Md, Andriananjamanantsoa & al. 00203387 (MT), KT826790*, KT826860*, KT826928*, –; Angraecum cf. zeylanicum Lindl., Sey, unvouchered (K), KF557956, KF558137, KF558230, –; Angraecum chaetopodium Schltr., Md, Andriananjamanantsoa & al. 00203462 (MT), KT826786*, KT826856*, KT826924*, ANGMD356-11; Angraecum clareae Hermans, Md, Hermans 3788 (K), KF557917, KF558049, KF558256, –; Angraecum compactum Schltr., Md, Andriananjamanantsoa & al. 00203420 (MT), KT826796*, KT826866*, KT826934*, –; Angraecum conchiferum Lindl., Africa, Bytebier 616 (EA), DQ091414, –, DQ091539, DQ091748; Angraecum conchoglossum Schltr., Md, Micheneau 5 (REU), KF558027, –, KF558242, –; Angraecum cordemoyi† Schltr., Md, Andriananjamanantsoa & al. 00203360 (MT), KT826782*, KT826852*, KT826920*, –; Angraecum cornigerum† Cordem., Md, Andriananjamanantsoa & al. 00203340 (MT), KT826809*, KT826879*, KT826947*, –; Angraecum corrugatum (Cordem.) Micheneau, Reu, Pailler 106 (REU), KF558003, KF558106, KF558194, DQ091745 (Carlsward 391, FLAS); Angraecum costatum Frapp. ex Cordem., Reu, Pailler 174 (REU), KF557942, KF558111, KF558180, –; Angraecum cucullatum Thouars, Reu, Pailler 108 (REU), KF557927, KF558051, KF558245, –; Angraecum danguyanum† H.Perrier, Md, Andriananjamanantsoa & al. 00203467 (MT), KT826818*, KT826887*, KT826956*, ANGMD305-11; Angraecum dasycarpum† Schltr., Md, Andriananjamanantsoa & al. 00203450 (MT), KT826817*, KT826886*, KT826955*, –; Angraecum didieri (Baill. ex Finet) Schltr., Md, unvouchered (K), KF558029, KF558030, KF558213, –; Angraecum dives Rolfe, Africa, Marimoto 42 (EA), DQ091422, –, DQ091547, DQ091756; Angraecum drouhardii† H.Perrier, Md, Andriananjamanantsoa & al. 00203424 (MT), KT826799*, KT826869*, KT826937*, ANGMD044-11; Angraecum dryadum† Schltr., Md, Andriananjamanantsoa & al. 00203341 (MT), KT826810*, KT826880*, KT826948*, –; Angraecum eburneum var. eburneum Bory, Reu, unvouchered (K), KF558000, KF558128, KF558261, DQ091742 (Carlsward 182, FLAS); Angraecum eburneum var. superbum† (Thouars) H.Perrier, Md, Andriananjamanantsoa & al. 00203463 (MT), KT826769*, KT826839*, KT826907*, DQ091738 (Carlsward 282, FLAS); Angraecum eburneum var. xerophilum H.Perrier, Md, unvouchered (K), KF557935, KF558092, KF558253, –; Angraecum elephantinum Schltr., Md, Carlsward 251 (FLAS), DQ091424, –, DQ091549, DQ091751; Angraecum expansum Thouars, Reu, Micheneau 2 (REU), KF557974, KF558087, KF558151, –; Angraecum filicornu† Thouars, Md, Andriananjamanantsoa & al. 00203460 (MT), KT826823*, KT826892*, KT826961*, ANGMD348-11; Angraecum florulentum Rchb.f., Com, unvouchered (K), KF557934, KF558088, KF558248, DQ091741 (Carlsward 321, FLAS); Angraecum germinyanum Hook.f., Md, Andriananjamanantsoa & al. 00203469 (MT), KT826765*, KT826835*, KT826904*, ANGMD254-11; Angraecum hermannii

iii

(Cordem.) Schltr., Md, unvouchered (REU), KF557970, KF558072, KF558159, –; Angraecum huntleyoides Schltr., Md, Hermans 4248 (K), KF557961, KF558143, KF558191, –; Angraecum lecomtei H.Perrier, Md, Andriananjamanantsoa & al. 00203440 (MT), KT826800*, KT826870*, KT826938*, ANGMD177-11; Angraecum leonis (Rchb.f.) Andre, Md, unvouchered (REU), KF557999, KF558116, KF558187, –; Angraecum letouzeyi Bosser, Md, Andriananjamanantsoa & al. 00203429 (MT), KT826804*, KT826874*, KT826942*, –; Angraecum liliodorum Frapp. ex Cordem., Reu, unvouchered (REU), KF557982, KF558100, KF558153, –; Angraecum linearifolium Garay, Md, Andriananjamanantsoa & al. 00203439 (MT), KT826767*, KT826837*, KT826906*, KF672215 (Simo-Droissart & al. 2152, BRLU); Angraecum longicalcar (Bosser) Senghas, Md, unvouchered (K), KF558013, KF558052, KF558210, DQ091739; Angraecum madagascariense (Finet) Schltr., Md, Andriananjamanantsoa & al. 00203448 (MT), KT826775*, KT826844*, KT826912*, –; Angraecum magdalenae Schltr. & H.Perrier, Md, unvouchered (K), KF557995, KF558123, KF558219, –; Angraecum mauritianum† (Poir.) Frapp., Md, Andriananjamanantsoa & al. 00203422 (MT), KT826815*, KT826884*, KT826953*, ANGMD039-11; Angraecum mirabile Schltr., Md, Andriananjamanantsoa & al. 00203465 (MT), KT826766*, KT826836*, KT826905*, ANGMD374-11; Angraecum moratii Bosser, Md, Andriananjamanantsoa & al. 00203447 (MT), KT826770*, KT826840*, KT826908*, –; Angraecum multiflorum Thouars, Reu, Pailler 154 (REU), KF557948, KF558053, KF558240, –; Angraecum musculiferum H.Perrier, Md, Andriananjamanantsoa & al. 00203432 (MT), KT826780*, KT826850*, KT826918*, –; Angraecum obesum H.Perrier, Md, Hermans 2407 (K), KF558011, KF558054, KF558249, –; Angraecum oblongifolium Toill.Gen. & Bosser, Md, Andriananjamanantsoa & al. 00203446 (MT), KT826771*, KT826841*, KT826909*, –; Angraecum obversifolium Frapp. ex Cordem., Reu, Pailler 8 (REU), KF557962, KF558125, KF558259; Angraecum ochraceum (Ridl.) Schltr., Md, Andriananjamanantsoa & al. 00203461 (MT), KT826785*, KT826855*, KT826923*, ANGMD354- 11; Angraecum panicifolium H.Perrier, Md, Andriananjamanantsoa & al. 00203444 (MT), KT826821*, KT826890*, KT826959*, ANGMD219-11; Angraecum pauciramosum Schltr., Md, Andriananjamanantsoa & al. 00203445 (MT), KT826781*, KT826851*, KT826919*, –; Angraecum pectinatum† Thouars, Md, Andriananjamanantsoa & al. 00203473 (MT), KT826819*, KT826888*, KT826957*, KF672211 (Simo-Droissart & al. 1684, BRLU); Angraecum penzigianum Schltr., Md, Andriananjamanantsoa & al. 00203466 (MT), KT826772*, KT826842*, KT826910*, –; Angraecum perparvulum† H.Perrier, Md, Andriananjamanantsoa & al. 00203326 (MT), KT826764*, KT826834*, KT826903*, –; Angraecum peyrotii Bosser, Md, Andriananjamanantsoa & al. 00203427 (MT), KT826813*, KT826882*, KT826951*, –; Angraecum pingue Frapp. ex Cordem.,

Md, Andriananjamanantsoa & al. 00203451 (MT), KT826792*, KT826862*, KT826930*, –; Angraecum pinifolium Bosser, Md, Andriananjamanantsoa & al. 00203425 (MT), KT826783*, KT826853*, KT826921*, –; Angraecum praestans Schltr., Md, Andriananjamanantsoa & al. 00203419 (MT), KT826814*, KT826883*, KT826952*, ANGMD074-11; Angraecum pseudodidieri H.Perrier, Md, unvouchered (K), KF558002, KF558032, KF558200, –; Angraecum pseudofilicornu† H.Perrier, Md, Andriananjamanantsoa & al. 00203435 (MT), KT826768*, KT826838*, –, –; Angraecum pterophyllum† H.Perrier, Md, Andriananjamanantsoa & al. 00203443 (MT), KT826761*, KT826832*, KT826901*, ANGMD218-11; Angraecum ramosum Thouars, Reu, Pailler 73 (REU), –, KF558037, KF558170, –; Angraecum rhynchoglossum† Schltr., Md, Andriananjamanantsoa & al. 00203457 (MT), KT826763*, –, –, –; Angraecum sacciferum Lindl., Africa, Bytebier 2226 (NBG), KF558028, KF558101, KF558181, –; Angraecum scottianum Rchb.f., Com, unvouchered (K), KF558017, KF558118, KF558163, AB217521 (TBG102594); Angraecum sedifolium† Schltr., Md, Andriananjamanantsoa & al. 00203396 (MT), KT826793*, KT826863*, KT826931*, –; Angraecum serpens (H.Perrier) Bosser, Md, Andriananjamanantsoa & al. 00203438 (MT), KT826773*, KT826843*, KT826911*, –; Angraecum sesquipedale Thouars, Md, Andriananjamanantsoa & al. 00203454 (MT), KT826824*, KT826893*, KT826962*, –; Angraecum setipes† Schltr., Md, Andriananjamanantsoa & al. 00203358 (MT), KT826788*, KT826858*, KT826926*, ANGMD358-11; Angraecum sororium† Schltr., Md, Andriananjamanantsoa & al. 00203400 (MT), KT826825*, KT826894*, KT826963*, –; Angraecum sterophyllum† Schltr., Md, Andriananjamanantsoa & al. 00203398 (MT), KT826812*, –, KT826950*, –; Angraecum striatum Thouars, Reu, Micheneau 4 (REU), KF557991, KF558058, KF558214, –; Angraecum tenuifolium Frapp. ex Cordem., Reu, Pailler 116 (REU), KF557968, KF558035, KF558254, –; Angraecum tenuispica Schltr., Md, unvouchered (K), KF557950, KF558099, KF558189, –; Angraecum teres Summerh., Africa, Bytebier 673 (EA), KF557969, KF558038, KF558198, DQ091755; Angraecum triangulifolium Senghas, Md, Andriananjamanantsoa & al. 00203442 (MT), KT826794*, KT826864*, KT826932*, –; Angraecum viguieri Schltr., Md, Andriananjamanantsoa & al. 00203423 (MT), KT826795*, KT826865*, KT826933*, –; Angraecum nanum† Frapp. ex Cordem., Md, Andriananjamanantsoa & al. 00203325 (MT), KT826797*, KT826868*, KT826936*, –; Beclardia macrostachya† (Thouars) A.Rich., Md, Andriananjamanantsoa & al. 00203322 (MT), KT826827*, KT826896*, –, –; Bolusiella iridifolia (Rolfe) Schltr., Africa, Bytebier 1113 (EA), DQ091356, –, DQ091481, DQ091665; Bolusiella maudiae (Bolus) Schltr., Africa, Bytebier 485 (EA), DQ091355, –, DQ091480, DQ091664; Calyptrochilum christyanum (Rchb.f.) Summerh., Africa, Carlsward 194 (SEL), DQ091325, –, DQ091449, DQ091668; Campylocentrum fasciola

(Lindl.) Cogn., America, Carlsward 301 (FLAS), DQ091321, –, DQ091445, DQ091564; Cribbia brachyceras (Summerh.) Senghas, Africa, Bytebier 361 (EA), DQ091365, –, DQ091490, DQ091577; Cribbia confusa P.J.Cribb, Africa, Kew 3936 (K), DQ091366, –, DQ091491, DQ091578; Cryptopus elatus (Thouars) Lindl., Reu, Micheneau 6 (REU), KF558019, KF558063, KF558221, DQ091585 (Carlsward 403, FLAS); Cryptopus paniculatus H.Perrier, Md, Hermans 5392 (K), KF557963, KF558113, KF558264, DQ091588 (Kew 5392, K); Cyrtorchis arcuata (Lindl.) Schltr., Africa, Bytebier 676 (EA), DQ091380, –, DQ091505, DQ091624; Cyrtorchis chailluana (Hook.f.) Schltr., Africa, Carlsward 156 (SEL), DQ091381, –, DQ091506, DQ091625; Dendrophylax funalis (Sw.) Benth. Ex Rolfe, America, Carlsward 302 (FLAS), KF557981, KF558067, KF558238, AF506310 (Whitten 1935, FLAS); Diaphananthe pellucida (Lindl.) Schltr., Africa, Carlsward 241 (FLAS), DQ091377, –, DQ091502, DQ091620; Diaphananthe sarcophylla (Schltr. ex Prain) P.J.Cribb & Carlsward, Africa, Bytebier 339 (EA), DQ091378, –, DQ091503, DQ091621; Diaphananthe vesicata (Lindl.) P.J.Cribb & Carlsward, Africa, Chase 14646 (K), KF557976, KF558083, KF558205, DQ091623 (Kew 399, K); Dolabrifolia bancoensis (Burg) Szlach. & Romowicz, Africa, Simo- Droissart & al. 54 (BRLU), KF672280, KF672257, KF662335, –; Dolabrifolia cf. aporoides (Summerh.) Szlach., Africa, Simo-Droissart & al. 15 (BRLU), KF672266, KF672230, KF662332, –; Dolabrifolia disticha (Lindl.) Szlach., Africa, Simo-Droissart & al. 6 (BRLU), KF672265, KF672231, KF662348, –; Dolabrifolia podochiloides (Schltr.) Szlach., Africa, Simo-Droissart & al. 12 (BRLU), KF672281, KF672238, KF662330, –; Eichlerangraecum birrimense (Rolfe) Szlach., Africa, Smith 531 (K), KF557930, KF558098, KF558196, –; Eichlerangraecum cf. eichlerianum (Kraenzl.) Szlach., Africa, unvouchered (K), KF557978, KF558069, KF558263, AF506322 (Carlsward 284, FLAS); Eichlerangraecum infundibulare (Lindl.) Szlach., Africa, unvouchered (K), KF557977, KF558077, KF558243, –; Erasanthe henricii (Schltr.) P.J.Cribb, Hermans & D.L.Roberts, Africa, unvouchered (K), –, KF558102, KF558175, –; Eurychone galeandrae (Rchb.f.) Schltr., Africa, Carlsward 293 (FLAS), DQ091349, –, DQ091473, DQ091614; Eurychone rothschildiana (O'Brien) Schltr., Africa, Carlsward 407 (FLAS), DQ091350, –, DQ091474, DQ091615; Jumellea anjouanensis (Finet) H.Perrier, Com, Bryonnaud 62 (REU), –, KF558120, KF558152, –; Jumellea arachnantha (Rchb.f.) Schltr., Md, Rakotoarivelo & al. 042 (REU), JQ905335, JQ905457, JQ905518, JQ905396; Jumellea arborescens H.Perrier, Md, unvouchered (K), KF557949, KF558095, KF558262, –; Jumellea bosseri Pailler, Reu, Pailler 270 (REU), JQ905340, JQ905462, JQ905523; Jumellea brevifolia H.Perrier, Md, Rakotoarivelo & al. 300 (TAN), JQ905342, JQ905464, JQ905525, JQ905401; Jumellea densifoliata Senghas, Md, Hermans 2809 (K), KF557966, KF558041, –, JQ905405 (Rakotoarivelo & al. 109, TAN); Jumellea exilis (Cordem.)

Schltr., Reu, Pailler 25 (REU), KF558014, KF558048, KF558169, –; Jumellea francoisii Schltr., Md, Rakotoarivelo & al. 230 (TAN), JQ905350, JQ905472, JQ905533, JQ905411; Jumellea hyalina H.Perrier, Md, Rakotoarivelo & al. 006 (TAN), JQ905354, JQ905476, JQ905537, JQ905415; Jumellea jumelleana (Schltr.) Summerh., Md, Rakotoarivelo & al. 099 (TAN), JQ905356, JQ905478, JQ905539, JQ905417; Jumellea longivaginans H.Perrier, Md, Rakotoarivelo & al. 328 (TAN), JQ905362, JQ905484, JQ905545, JQ905423; Jumellea majalis (Schltr.) Schltr., Md, Rakotoarivelo & al. 307 (TAN), JQ905366, JQ905488, JQ905549, JQ905427; Jumellea maxillarioides (Ridl.) Schltr., Md, unvouchered (K), KF558004, KF558134, KF558161, –; Jumellea rigida Schltr., Md, Rakotoarivelo & al. 220 (TAN), JQ905376, JQ905498, JQ905559, JQ905437; Jumellea spathulata (Ridl.) Schltr., Md, Rakotoarivelo & al. 209 (TAN), JQ905380, JQ905502, JQ905563, JQ905441; Jumellea tenuibracteata (H.Perrier ex Hermans) F.P.Rakotoar. & Pailler, Md, Rakotoarivelo & al. 321 (TAN), JQ905360, JQ905482, JQ905543, –; Jumellea teretifolia Schltr., Md, Rakotoarivelo & al. 160 (TAN), JQ905386, JQ905508, JQ905568, JQ905447; Lemurella pallidiflora Bosser, Md, Andriananjamanantsoa & al. 00203431 (MT), KT826828*, KT826897*, KT826965*, –; Lemurella papillosa Bosser, Md, Andriananjamanantsoa & al. 00203434 (MT), KT826829*, KT826898*, KT826966*, –; Lemurorchis madagascariensis Kraenzl., Md, Kew 5383 (K), DQ091431, –, DQ091556, DQ091747; Listrostachys pertusa (Lindl.) Rchb.f., Africa, Carlsward 399 (FLAS), DQ091384, –, DQ091509, DQ091637; Microcoelia gilpinae (Rchb.f. & Moore) Summerh., Md, Carlsward 290 (FLAS), DQ091397, –, DQ091522, DQ091649; Microcoelia stolzii (Schltr.) Summerh., Africa, Carlsward 287 (FLAS), DQ091405, –, DQ091530, DQ091657; Mystacidium capense (L.f.) Schltr., Africa, Whitten 1781 (FLAS), DQ091362, –, DQ091487, DQ091573; Mystacidium flanaganii (Bolus) Bolus, Africa, Kew 5084 (K), DQ091363, –, DQ091488, DQ091574; Neobathiea grandidierana (Rchb.f.) Garay, Md, Carlsward 395 (FLAS), DQ091329, –, DQ091453, DQ091589; Oeonia rosea† Ridl., Md, Andriananjamanantsoa & al. 00203385 (MT), KT826830*, KT826899*, –, DQ091736 (Carlsward 221, FLAS); Oeoniella polystachys† (Thouars) Schltr., Md, Andriananjamanantsoa & al. 00203455 (MT), KT826831*, KT826900*, –, –; Pectinariella atlantica Stevart & Droissart, Africa, Simo-Droissart & al. 2065 (BRLU), KF672274, KF672235, KF662349, – ; Pectinariella doratophylla (Summerh.) Szlach., Africa, Simo-Droissart & al. 1466 (BRLU), KF672261, KF672247, KF662351, –; Pectinariella gabonensis (Summerh.) Szlach., Africa, Simo- Droissart & al. 1468 (BRLU), KF672279, KF672237, KF662333, –; Pectinariella pungens (Schltr.) Szlach., Africa, Simo-Droissart & al. 1464 (BRLU), KF672278, KF672254, KF662329, –; Pectinariella subulata (Lindl.) Szlach., Africa, Simo-Droissart & al. 1474 (BRLU), KF672285, KF672251, KF662355, –; Phalaenopsis cornu-cervi (Breda) Blume & Rchb.f., Asia, Chase O-1356

vii

(K), KF558008, KF558082, –, –; Podangis dactyloceras (Rchb.f.) Schltr., Africa, Kew 4999 (K), DQ091385, –, DQ091510, DQ091628; Polystachya fulvilabia Schltr., Africa, Chase 17862 (K), KF558010, KF558057, KF558235, GU556680; Rangaeris amanuensis (Kraenzl.) Summerh., Africa, Bytebier & Kirika 26 (EA), DQ091386, –, DQ091512, –; Rangaeris muscicola (Rchb.f.) Summerh., Africa, Carlsward 169 (SEL), DQ091387, –, DQ091513, DQ091630; Rhipidoglossum subsimplex (Summerh.) Garay, Africa, Bytebier 546 (EA), DQ091371, –, DQ091496, DQ091580; Rhipidoglossum xanthopollinium (Rchb.f.) Schltr., Africa, Carlsward 384 (FLAS), DQ091370, –, DQ091495, DQ091582; Sobennikoffia humbertiana H.Perrier, Md, Kew 3044 (K), DQ091433, –, DQ091558, DQ091750; Solenangis clavata (Rolfe) Schltr., Africa, Carlsward 397 (FLAS), DQ091409, –, DQ091534, DQ091666; Solenangis wakefieldii (Rolfe) P.J.Cribb & J.Stewart, Africa, Bytebier 627 (EA), DQ091410, –, DQ091535, DQ091667; Sphyrarhynchus schliebenii Mansf., Africa, Bytebier 393 (EA), DQ091359, –, DQ091484, –; Tridactyle bicaudata (Lindl.) Schltr., Africa, Bytebier 348 (EA), DQ091388, –, DQ091514, DQ091638; Tridactyle filifolia (Schltr.) Schltr., Africa, Bytebier 707 (EA), DQ091390, –, DQ091516, DQ091641; Vanda tricolor Lindl., Africa, Chase 17970 (K), KC823021, –, KC985407, –; Ypsilopus longifolius (Kraenzl.) Summerh., Africa, Bytebier 609 (EA), DQ091393, –, DQ091519, DQ091636; Ypsilopus viridiflorus P.J.Cribb & J.Stewart, Africa, Bytebier 402 (EA), DQ091395, –, –, DQ091633.

viii

Annexe 2. Morphological character description

Description of vegetative and floral morphological characters in Angraecinae N° Characters Character states Description of states References 1 habitat 1 terrestrial growing on the ground Beentje, 2010 2 lithophytic growing on rock Beentje, 2010 3 epiphytic growing on and attached to Beentje, 2010 another plant without deriving nourishment from it

2 habit 1 pendent hanging Beentje, 2010 2 horizontal parallel to the substrate 3 vertical perpendicular to the substrate 4 erect uprigth Beentje, 2010 5 climber growing upwards by Beentje, 2010 attaching itself to other structures which It uses as supports (tree or rock)

3 natural spread 1 tiny 0 cm to 3 cm (height) 2 small 3 cm to 5 cm 3 medium 5 cm to 15 cm 4 large 15 cm to 35 5 very-large > 35 cm 6 N/A not applicable 4 natural spread 1 tiny 0 cm to 3 cm (width) 2 small 3 cm to 5 cm 3 medium 5 cm to 15 cm 4 large 15 cm to 35 5 very-large > 35 cm 6 N/A not applicable 5 internode 1 very-short L < 0.3 cm 2 short 1 cm < L < 0.4 cm 3 medium 4 cm < L < 2 cm 4 long L > 5 cm 5 N/A not applicable 6 diameter of 1 tiny d < 0.2 cm stem 2 small 0.5 cm < d < 0.3 cm 3 medium 1.2 cm < d < 0.6 cm 4 large d > 1.3 cm 5 N/A not applicable

7 form of leaf 1 acicular needle-shaped; very narrow, Beentje, 2010 stiff, and pointed 2 cylindrical long and narrow with Beentje, 2010 circular cross-section 3 linear narrow and much longer Beentje, 2010 than wide, with parallel margins 4 lanceolate narrowly ovate and tapering Beentje, 2010 to a point at the apex 5 oblong longer than broad, with the Beentje, 2010 margins parallel for most of their length (1.5 - 2 x as long as wide)

6 elliptic broadest at the middle with Beentje, 2010 two equal rounded ends, width ratio 1.5 - 2 7 triangular 8 N/A not applicable 8 apex of leaf 1 obtuse not pointed, blunt, ending in Beentje, 2010 an angle of between 90 - 180° 2 attenuate gradually narrowing over a Beentje, 2010 long distance 3 emarginate with a distinct sharp notch, Beentje, 2010 both ends not at the same level 4 incised having straight to irregular Kiger & Porter, lines of separation extending 2001 inward from the margin

5 N/A not applicable 9 texture of leaf 1 herbaceous with the texture of a herb, Beentje, 2010 soft and pliable 2 membraneous like a membran, flexible and Beentje, 2010 thin 3 fleshy swollen largely because of Beentje, 2010 high water content 4 succulent thick, fleshy and swollen Beentje, 2010 5 coriaceous moderately thick, and tough Kiger & Porter, 2001 6 fibrous composed of fibers Beentje, 2010 7 N/A not applicable

10 leaf-surface 1 smooth even or unrelieved overall Kiger & Porter, 2001 2 rugose covered in reticulate lines, Beentje, 2010 with the spaces in between convex 3 N/A not applicable 11 numbre of 1 few 0 to 5 leaves 2 medium 5 to 10 3 many 10 to 25 4 dense > 25 5 N/A not applicable 12 leaf-width 1 filiform 0 to 2 mm 2 very-narrow 3 to 5 mm 3 narrow 6 to 10 mm 4 medium 11 to 25 mm 5 large 26 to 40 mm 6 N/A not applicable 13 leaf-length 1 very-short 0 to 3 cm 2 short 3 to 7 cm 3 medium 7 to 15 cm 4 long 15 to 30 cm 5 very-long > 30 cm 6 N/A not applicable 14 type of 1 solitary type 1 1 to 3 solitary flowers inflorescence 2 solitary type 2 more than 3 solitary flowers 3 raceme type 1 1 to 3 raceme bearing up to 5 flowers each 4 raceme type 2 more than 3 racemes bearing up to 5 flowers each 5 raceme type 3 1 to 3 raceme bearing more than 5 flowers each 6 raceme type 4 more than 3 racemes bearing more than 5 flowers each

7 N/A 15 length of 1 short almost sessile inflorescence 2 medium 1/3 of the size of the plant 3 long 1/2 of the size of the plant 4 very-long longer than the plant 16 length of 1 short 1/4 of the length of ovary pedicel 2 medium 1/2 of the length of ovary 3 long same length as the ovary

4 very-long longer than the ovary 17 position of 1 terminal ending the axis Beentje, 2010 inflorescence 2 subterminal lateral and just below the Kiger & Porter, apex 2001 3 suprafoliar upon the stem directly above Kiger & Porter, a leaf insertion 2001 4 infrafoliar upon the stem directly below Kiger & Porter, a leaf insertion 2001 5 N/A not applicable 18 size of flower 1 tiny labellum (L) less than 0.3 cm (spur excluded) 2 small 0.4 cm < L < 0.9 cm 3 medium 2 cm < L < 1.9 cm 4 large 5 cm < L < 2 cm 5 very-large L > 5 cm 19 form of sepal 1 ovate egg shaped (2-dimensional), Beentje, 2010 about 1.5x as long as broad, with the wider part below the middle

2 linear 3 lanceolate 4 tubular cylindrical and hollow Beentje, 2010 5 obovate egg-shaped with the Beentje, 2010 broadest part near the apex

6 elliptic 20 color of sepal 1 green to yellowish 2 white to white green 3 ocher 4 other 21 apex of sepal 1 acute sharp, sharply pointed, the Beentje, 2010 margins near the tip being almost straight and forming an angle of < 90°

2 attenuate 3 obtuse 22 form of petal 1 ovate 2 linear 3 lanceolate

xii

4 tubular 5 obovate 6 elliptic 7 N/A 23 color of petal 1 green to yellowish 2 white to white green 3 ocher 4 other 24 apex of petal 1 acute 2 attenuate 3 obtuse 4 lobed 25 orientation of 1 straight dorsal sepal 2 inflexed bent or curved inwards Beentje, 2010 3 reflexed curved backwards Beentje, 2010 26 orientation of 1 straight lateral sepal 2 inflexed 3 reflexed 27 orientation of 1 straight lateral petal 2 inflexed 3 reflexed 28 2D form of 1 lanceolate labellum 2 ovate 3 orbicular 4 elliptic 5 rhombic 6 trullate 7 obovate 8 linear 29 3D form of the 1 plane labellum 2 scoop-shaped slightly concave from the type 1 base to the tip 3 scoop-shaped very concave from the base type 2 to the tip 4 navicular 5 infundibular 6 gibbous 30 position of 0 uppermost labellum 1 lowermost

xiii

31 color of 1 greenish to labellum yellowish 2 white to white green 3 other 32 apex of 1 round/obtuse labellum 2 acute 3 attenuate 4 caudate 5 lobed 6 waved 7 dentate 33 texture of 1 thin flower 2 thick 3 fleshy 4 succulent 34 spur-length 1 very-short L < 0.5 cm 2 short L < 3 cm 3 medium L < 7 cm 4 long L < 15 cm 5 very-long L > 15 cm 35 form of spur 1 straight 2 arcute type 1 arched at median (bow-like) 3 arcute type 2 arched at the base (club-like)

4 arcute type 3 very arched at median (swan neck-like) 5 sigmoid 6 circinnate 7 N/A not applicable 36 base of spur 1 isodiametric same diameter form the base to the tip 2 tapering type 1 large at the base, and become gradually isodiametric towards the tip 3 tapering type 2 very large at the base, and become abruptley isodiametric towards the tip 4 tapering type 3 very large at the base, and become narrow gradually towards the tip 5 N/A not applicable

xiv

37 color of spur 1 green to yellowish 2 white to white green 3 ocher 4 other 5 N/A not applicable 38 apex of spur 1 acute 2 obtuse 3 N/A not applicable 39 orientation of 1 straight lacking significant curves or Kiger & Porter, spur bends 2001 2 ascending spreading at the base and Kiger & Porter, then curving upward or 2001 forward, the distal portion more or less parallel to the bearing structure

3 descending salient at its base and then Kiger & Porter, curving downward or 2001 backward 4 N/A not applicable

References:

Beentje, H.J. 2010. The Kew plant glossary: an illustrated dictionary of plant terms. Royal Botanic Gardens, Kew, 164 pages. Kiger, R.W. & Porter, D.M. 2001. Categorical glossary for the flora of North America project. Pittsburgh, 165 pages.

Annexe 3. Morphological character matrix

Morphological character matrix used in the phylogenetic reconstruction.

Taxa 123456789101112131415161718192021222324 252627282930313233343536373839 Acampe ochracea 335433535134362112543542222712313175334 Aerangis ellisii 235443615135454223321322333712212421223 Aerangis hariotiana 324413633125364131331322333112324111223 Aerangis hildebrandtii 323313633125364131331322333112324111223 Aerangis macrocentra 323412613115354233321322333712212321223 Aerangis punctata 322312613115334233321322333712212421223 Aeranthes aemula 334413332124414133112112222262131213123 Aeranthes arachnites 334413332124414133112112222262131213123 Aeranthes caudata 334413332124414133112112222262131213123 Aeranthes grandiflora 334413332124414133112112222262131213123 Aeranthes longipes 334413332124414133112112222262131213123 Aeranthes moratii 334413332124414133112112222262131213123 Aeranthes neoperrieri 334413332124414133112112222262131213123 Aeranthes strangulata 334413332124414133112112222262131213123 Aeranthes tenella 334413332124414133112112222262131213123 Aeranthes virginalis 334413332124414133112112222262131213123 Aerides odorata 345433335145464214341342111412352264323 Ancistrorhynchuscapitatus 315413335124471132623622222222214221223 Ancistrorhynchusrecurvus 315413335124571132623622222222214251123 Angraecoides cf. moandense 345342435143222123312312322111133211112 Angraecoides chevalieri 345342435143222123312312322111133211112 Angraecoides cultriformis 345342445133222123312312322111133211112 Angraecoides erecta 255342445143222123312312322111133211112 Angraecopsis parviflora 322312423124254132311312333511154211123 Angraecum amplexicaule 344332421133211413321321322441232451212 Angraecum ankeranense 343223635223111324321321121212222461213 Angraecum appendiculatum 344242635143112313432422333111232111112 Angraecum arachnites 344232632132122313432422333331242452212

xvi

Angraecum baronii 315322423242121122311311322442134111122 Angraecum borbonicum 333313535114311224321321111212222421213 Angraecum bracteosum 334524335125462123121121222222222211123 Angraecum cadetii 334524335125463123121121222222222211123 Angraecum calceolus 343422421124464232311311333441134211122 Angraecum caricifolium 315232332142263121311311222442134111121 Angraecum caulescens 343422421124464232311311333441134211121 Angraecum cf. breve 332313535213211323321321111212222421213 Angraecum cf. elephantinum 344423535114311124321321121212222421213 Angraecum cf. germinyanum 344222632143122313432422333331242452212 Angraecum cf. humblotianum 343111225241111112321321222212222111221 Angraecum cf. humile 311111323122131121121121222251222111121 Angraecum cf. liliodorum 333313535213111124321321121212222421213 Angraecum cf. panicifolium 243331431221211112321321222212222111221 Angraecum cf. pauciramosum 315332331143263121311311222442134111121 Angraecum cf. peyrotii 333313535113111124321321121212222421213 Angraecum cf. pinifolium 343322421121333232311311322442134111121 Angraecum cf. rutenbergianum 233313535123211124321321121212222421213 Angraecum cf. sacculatum 343312421133343121311311322442134111121 Angraecum cf. zeylanicum 343422421124464132311311333441134211122 Angraecum chaetopodium 323312421123214132311311333441134211122 Angraecum clareae 344424533234231325321322123232233443213 Angraecum compactum 344424533224331325321322123232223443213 Angraecum conchiferum 343342335222112313432422333431242352212 Angraecum conchoglossum 344222632133122313432422333331242452212 Angraecum cordemoyi 323312421123232122311311322442134111122 Angraecum cornigerum 344342235222212424321321121212222421213 Angraecum corrugatum 344332635144212313432422333111232111112 Angraecum costatum 355332421142121122311312322442134111122 Angraecum cucullatum 332213635122111323321322111212222414213 Angraecum danguyanum 243331421131312213321322222412212321213 Angraecum dasycarpum 243222615133111112321321222212222111221 Angraecum didieri 343223635223111324321322121212222461213 Angraecum dives 334413325134364131311312222142132111121

xvii

Angraecum drouhardii 332213635222111323321322111212222261213 Angraecum dryadum 343223635223111324321321121212222461213 Angraecum eburneum var. eburneum 235524335145564125311312111321243311111 Angraecum eburneum var. superbum 335524335145564125311312111321243411111 Angraecum eburneum var. xerophilum 235534335144464125311312111321243311111 Angraecum elephantinum 344423535124311124321322121212222421213 Angraecum expansum 344342635223112313432422333431242352212 Angraecum filicornu 244331431122312213321322222212222421213 Angraecum florulentum 315332435244232323321322322441232451212 Angraecum germinyanum 344222632143122313432422333331242452212 Angraecum hermannii 315322423242121122311312322442134111122 Angraecum huntleyoides 335514421134564432311312333441134211112 Angraecum lecomtei 332312235221111323321321111112222461213 Angraecum leonis 333414523115432325321322121232223443213 Angraecum letouzeyi 333312226221312323321322121212222321213 Angraecum liliodorum 333313535213111124321322121212222421213 Angraecum linearifolium 343322126131313313432422333431242452212 Angraecum longicalcar 235524335145564125311312111321243511111 Angraecum madagascariense 314122622143142121311311222442134111121 Angraecum magdalenae 234524535125331225321322121532233443213 Angraecum mauritianum 244322422134212213321322222212222421213 Angraecum mirabile 343342335222112313432422333431242452212 Angraecum moratii 315332635144222323321322333441232451212 Angraecum multiflorum 343422421132263122311312333441134211121 Angraecum musculiferum 315232332143263121311312222442134111121 Angraecum nanum 331111421122144131311312222142132111121 Angraecum obesum 344523335124311124321322121212232421213 Angraecum oblongifolium 314222635133111113321322333441232351212 Angraecum obversifolium 343312421133333122311312322442134111122 Angraecum ochraceum 333312421123314132311311333441124211122 Angraecum panicifolium 243331431221211112321322222212222111221 Angraecum pauciramosum 314132321142163121311312222442134111121 Angraecum pectinatum 243222535232111112321322222212222111221 Angraecum penzigianum 315432423144342323321322333441232321212

xviii

Angraecum perparvulum 321111613113131131121122222251221111121 Angraecum peyrotii 333313535113312124321322121212222421213 Angraecum pingue 314222724133122112212212123152132211122 Angraecum pinifolium 343322421121333232311312322442134111121 Angraecum praestans 234524335125433225212212111532242443213 Angraecum pseudodidieri 344323335123211124321322121212232421213 Angraecum pseudofilicornu 313322225122212214431422333331232452212 Angraecum pterophyllum 311111323222131121121122222251221111121 Angraecum ramosum 254232425242212213321322222212222421213 Angraecum rhynchoglossum 322211422114214132211212333151141211111 Angraecum sacciferum 321211421124244131311312322442134111121 Angraecum scottianum 314322225132312214431422333331232452212 Angraecum sedifolium 333111224142122112212212123152132211122 Angraecum serpens 315432423144312323321322333441232221212 Angraecum sesquipedale 235524335145433225321322111212223521113 Angraecum setipes 333312421123314132311312333441124211122 Angraecum sororium 235524535145432225321322111212223521113 Angraecum sterophyllum 343313535222213223321322121212222421213 Angraecum striatum 334524335125464123121122222222222211123 Angraecum tenuifolium 344332421122253222311312322442134111121 Angraecum tenuispica 314422421143263121311312222442134111121 Angraecum teres 333413325121364432311312222142132211121 Angraecum triangulifolium 314222724133122112212212123152132211122 Angraecum viguieri 345434332134322125232212121532242443213 Beclardia macrostachya 344423335125454234321522222752252213121 Bolusiella iridifolia 333413523122254131321322222112221121223 Bolusiella maudiae 333413523123154131321322222112221121223 Calyptrochilum christyanum 335444535134361122311312222712154212123 Campylocentrum fasciola 356655857356664151121122222232212111221 Cribbia brachyceras 344423335124364322211212222122124211123 Cribbia confusa 343423335124364323211212222122124211123 Cryptopus elatus 345432535124254123523722333812252113211 Cryptopus paniculatus 345432535124254123513712333812252113221 Cyrtorchis arcuata 315523635125464224221222333112222322113

xix

Cyrtorchis chailluana 315523635125464224221222333112222422113 Dendrophylax funalis 356655857356623353311312111332252313111 Diaphananthe pellucida 315413635125564123321322222212271221213 Diaphananthe sarcophylla 315423443224464121331332333112324221313 Diaphananthe vesicata 315423443224564121311312333112124221123 Dolabrifolia bancoensis 313111612143121122321321111441231111213 Dolabrifolia cf. aporoides 313111612144121122321321111441231111213 Dolabrifolia disticha 313111612143121122321322111441231111213 Dolabrifolia podochiloides 315111612142121122321322111441231111213 Eichlerangraecum birrimense 355332435145422225311311333752233314123 Eichlerangraecum cf. eichlerianum 355332435145422225311311333752233444123 Eichlerangraecum infundibulare 355332435135422225311312333752233314123 Erasanthe henricii 323413635125454135322322222762272411113 Eurychone galeandrae 323413633124454134231222222332371214222 Eurychone rothschildiana 323413633124453134221222222232271214222 Jumellea anjouanensis 335413335134411323221222123612222431113 Jumellea arachnanta 335413335134411323221222123612222431113 Jumellea arborescens 335413335134411323221222123612222431113 Jumellea bosseri 335413335134411323221222123612222431113 Jumellea brevifolia 335413335134411323221222123612222431113 Jumellea densifoliata 335413335134411323221222123612222431113 Jumellea exilis 335413335134411323221222123612222431113 Jumellea francoisii 335413335134411323221222123612222431113 Jumellea hyalina 335413335134411323221222123612222431113 Jumellea jumelleana 335413335134411323221222123612222431113 Jumellea longivaginans 335413335134411323221222123612222431113 Jumellea majalis 335413335134411323221222123612222431113 Jumellea maxillarioides 335413335134411323221222123612222431113 Jumellea rigida 335413335134411323221222123612222431113 Jumellea spathulata 335413335134411323221222123612222431113 Jumellea tenuibracteata 335413335134411323221222123612222431113 Jumellea teretifolia 335413335134411323221222123612222431113 Lemurella pallidiflora 343322423124234122311312333152154212112 Lemurella papillosa 343322423124234122311312333152154212112

Lemurorchis madagascariensis 335524335145564122331332111252354211213 Listrostachys pertusa 345514335144464122121122333712224111123 Microcoelia gilpinae 356655857356663452133132222222314214323 Microcoelia stolzii 356655857356664151623622222412211124223 Mystacidium capense 323412433114364233321322333112221311213 Mystacidium flanaganii 322312433112164232311312333112121211113 Neobathiea grandidierana 314333635124232214321322333112222443113 Oeonia rosea 355341625122133123513512111852252111221 Oeoniella polystachys 254332535124364223311312222352242111112 Pectinariella atlantica 315221425242221122321321111441231111213 Pectinariella doratophylla 315221425242221122321322111441231211213 Pectinariella gabonensis 315221425242221122321322111441231111213 Pectinariella pungens 315221425242221122321322111441231111213 Pectinariella subulata 315331225241221122321322111441231111213 Phalaenopsis cornu-cervi 322412613115364224341342111212322175334 Podangis dactyloceras 343313425124363422621622222422261224223 Polystachya fulvilabia 133312421122453342341342222221322175334 Rangaeris amaniensis 345514335124454323321322333512221421113 Rangaeris muscicola 344514335124354322321322333212221421213 Rhipidoglossum subsimplex 315343435133262122613112222712161211113 Rhipidoglossum xanthopollinium 315443435133363122613112222712161221113 Sobennikoffia humbertiana 234424335145564224321322323732252213122 Solenangis clavata 355342635134141222613612222112151221123 Solenangis wakefieldii 355342635134142222623622222112252324213 Sphyrarhynchus schliebenii 323313425113244332321322112222251221223 Tridactyle bicaudata 335444335124362122311212333512151211111 Tridactyle filifolia 315322225212341132311212333512151211111 Vanda tricolor 345433335144454315543542222712352175334 Ypsilopus longifolius 314313322123454332321322333512231321213 Ypsilopus viridiflorus 314313322123422132211212333112131221113

xxi

Annexe 4. Characters used in BAMM analyses

Continuous characters used for the phenotypic/evolution model, and fractioning used for the speciation/extinction model implemented in BAMM, values for flower size and spur length were log-transformed and fraction represents the percentage of samples used in the analyses for each clade (from 0 to 1). Clade Taxa Flower size Spur length Fraction Aeranthes Aeranthes aemula 2.255272505 1.342422681 0.3 Aeranthes Aeranthes arachnites 2.158362492 1.301029996 0.3 Aeranthes Aeranthes caudata 2.021189299 1.397940009 0.3 Aeranthes Aeranthes grandiflora 2.278753601 1.477121255 0.3 Aeranthes Aeranthes longipes 2.130333768 1.301029996 0.3 Aeranthes Aeranthes moratii 2.049218023 1.255272505 0.3 Aeranthes Aeranthes neoperrieri 2.130333768 1.342422681 0.3 Aeranthes Aeranthes strangulata 2.049218023 1.301029996 0.3 Aeranthes Aeranthes tenella 2.158362492 1.176091259 0.3 Aeranthes Aeranthes virginalis 2.049218023 1.342422681 0.3 Angraecum Angraecum amplexicaule 1.77815125 1.954242509 0.8 Angraecum Angraecum ankeranense 2.491361694 1.939519253 0.8 Angraecum Angraecum appendiculatum 2.380211242 -1.958607315 0.8 Angraecum Angraecum arachnites 1.77815125 2.041392685 0.8 Angraecum Angraecum baronii 0 0.698970004 0.8 Angraecum Angraecum cf. germinyanum 2.380211242 2.041392685 0.8 Angraecum Angraecum borbonicum 2.342422681 1.84509804 0.8 Angraecum Angraecum bracteosum 1.176091259 0.602059991 0.8 Angraecum Angraecum cadetii 1.322219295 0.698970004 0.8 Angraecum Angraecum calceolus 1.397940009 1.176091259 0.8 Angraecum Angraecum caricifolium 0.477121255 0.477121255 0.8 Angraecum Angraecum caulescens 0.954242509 1 0.8 Angraecum Angraecum cf. breve 1.681241237 2.079181246 0.8 Angraecum Angraecum cf. peyrotii 2.342422681 2.255272505 0.8 Angraecum Angraecum cf. elephantinum 2.397940009 2.041392685 0.8 Angraecum Angraecum cf. germinyanum 2.477121255 2.079181246 0.8 Angraecum Angraecum cf. pauciramosum 0 0 0.8 Angraecum Angraecum cf. humblotianum 1.544068044 0.477121255 0.8 Angraecum Angraecum cf. liliodorum 2.653212514 1.77815125 0.8 Angraecum Angraecum cf. panicifolium 1.544068044 2.004321374 0.8 Angraecum Angraecum cf. pinifolium 0.602059991 0.602059991 0.8 Angraecum Angraecum cf. rutenbergianum 2.498310554 2.146128036 0.8 Angraecum Angraecum cf. sacculatum 0.602059991 0.477121255 0.8

xxii

Angraecum Angraecum cf. zeylanicum 1.204119983 1 0.8 Angraecum Angraecum chaetopodium 0.698970004 1.342422681 0.8 Angraecum Angraecum ochraceum 0.698970004 1.342422681 0.8 Angraecum Angraecum clareae 2.45484486 1.079181246 0.8 Angraecum Angraecum compactum 2.318063335 2.041392685 0.8 Angraecum Angraecum conchiferum 1.924279286 1.698970004 0.8 Angraecum Angraecum conchoglossum 2.075546961 2.06069784 0.8 Angraecum Angraecum cordemoyi 0.954242509 0.602059991 0.8 Angraecum Angraecum cornigerum 2.290034611 2.113943352 0.8 Angraecum Angraecum corrugatum 2.021189299 -1.958607315 0.8 Angraecum Angraecum costatum 0.602059991 0.477121255 0.8 Angraecum Angraecum cucullatum 1.875061263 1.079181246 0.8 Angraecum Angraecum danguyanum 1.681241237 1.653212514 0.8 Angraecum Angraecum dasycarpum 0.954242509 0.602059991 0.8 Angraecum Angraecum didieri 2.167317335 1.929418926 0.8 Angraecum Angraecum dives 0.698970004 0.903089987 0.8 Angraecum Angraecum drouhardii 1.62324929 1.230448921 0.8 Angraecum Angraecum dryadum 1.963787827 1.954242509 0.8 Angraecum Angraecum ebu.eburneum 2.720159303 1.77815125 0.8 Angraecum Angraecum ebu.superbum 2.670245853 1.84509804 0.8 Angraecum Angraecum ebu.xerophilum 2.552668216 1.875061263 0.8 Angraecum Angraecum elephantinum 2.698970004 2.041392685 0.8 Angraecum Angraecum expansum 1.924279286 1.77815125 0.8 Angraecum Angraecum filicornu 1.342422681 1.939519253 0.8 Angraecum Angraecum florulentum 1.799340549 1.954242509 0.8 Angraecum Angraecum hermannii 0.477121255 0.301029996 0.8 Angraecum Angraecum huntleyoides 1.681241237 1.301029996 0.8 Angraecum Angraecum lecomtei 1.62324929 1.977723605 0.8 Angraecum Angraecum leonis 2.301029996 1.954242509 0.8 Angraecum Angraecum letouzeyi 1.431363764 2.041392685 0.8 Angraecum Angraecum liliodorum 2.653212514 1.77815125 0.8 Angraecum Angraecum linearifolium 1.954242509 2 0.8 Angraecum Angraecum longicalcar 2.684845362 2.431363764 0.8 Angraecum Angraecum madagascariense 0 0 0.8 Angraecum Angraecum magdalenae 2.498310554 1.986771734 0.8 Angraecum Angraecum mauritianum 1.84509804 1.903089987 0.8 Angraecum Angraecum mirabile 1.954242509 2.06069784 0.8 Angraecum Angraecum moratii 2.130333768 2.06069784 0.8 Angraecum Angraecum multiflorum 0.602059991 0.602059991 0.8 Angraecum Angraecum musculiferum 0.477121255 0.477121255 0.8 Angraecum Angraecum obesum 2.431363764 1.929418926 0.8 Angraecum Angraecum oblongifolium 1.707570176 1.763427994 0.8

xxiii

Angraecum Angraecum obversifolium 0.301029996 0.301029996 0.8 Angraecum Angraecum panicifolium 1.544068044 2.004321374 0.8 Angraecum Angraecum pauciramosum 0 0.301029996 0.8 Angraecum Angraecum pectinatum 1.113943352 0.301029996 0.8 Angraecum Angraecum penzigianum 1.755874856 1.662757832 0.8 Angraecum Angraecum peyrotii 2.45484486 1.986771734 0.8 Angraecum Angraecum pingue 0.84509804 1.544068044 0.8 Angraecum Angraecum pinifolium 0.602059991 0.602059991 0.8 Angraecum Angraecum praestans 2.737192643 1.986771734 0.8 Angraecum Angraecum pseudodidieri 2.155336037 1.740362689 0.8 Angraecum Angraecum pseudofilicornu 2.176091259 2.096910013 0.8 Angraecum Angraecum ramosum 1.903089987 1.544068044 0.8 Angraecum Angraecum sacciferum 0.301029996 0.301029996 0.8 Angraecum Angraecum scottianum 2.73239376 2.176091259 0.8 Angraecum Angraecum sedifolium 1.301029996 1.113943352 0.8 Angraecum Angraecum serpens 1.944482672 1.568201724 0.8 Angraecum Angraecum sesquipedale 2.812913357 2.431363764 0.8 Angraecum Angraecum setipes 0.602059991 0.954242509 0.8 Angraecum Angraecum sororium 3.021189299 2.505149978 0.8 Angraecum Angraecum sterophyllum 2.008600172 2 0.8 Angraecum Angraecum striatum 1.322219295 0.903089987 0.8 Angraecum Angraecum tenuifolium 0.477121255 0.602059991 0.8 Angraecum Angraecum tenuispica 0.217483944 0.301029996 0.8 Angraecum Angraecum teres 0.477121255 1.380211242 0.8 Angraecum Angraecum triangulifolium 1.255272505 1.397940009 0.8 Angraecum Angraecum viguieri 2.77815125 2.079181246 0.8 Angraecum Angraecum nanum 0 0 0.8 Angraecum Angraecum zaratananae 1.602059991 1.544068044 0.8 Angraecum Oeoniella polystachys 2.397940009 0.698970004 0.8 Angraecum Sobennikoffia humbertiana 2.531478917 1.477121255 0.8 CladeIA Angraecum cf. humile 0.301029996 0.301029996 0.4 CladeIA Angraecum perparvulum 0 0 0.4 CladeIA Angraecum pterophyllum 0.77815125 0.477121255 0.4 CladeIA Angraecum rhynchoglossum 0.477121255 1.431363764 0.4 CladeIA Beclardia macrostachya 2.903089987 1.414973348 0.4 CladeIA Cryptopus elatus 2.62324929 0.698970004 0.4 CladeIA Cryptopus paniculatus 2.720159303 0.698970004 0.4 CladeIA Erasanthe henricii 2.829303773 2.176091259 0.4 CladeIA Lemurella pallidiflora 1.380211242 1.414973348 0.4 CladeIA Lemurella papillosa 1.380211242 1.380211242 0.4 CladeIA Neobathiea grandidierana 2.459392488 2.079181246 0.4 CladeIA Oeonia rosea 2.397940009 0.602059991 0.4

xxiv

CladeIB Ancistrorhynchus capitatus 1.204119983 1.414973348 0.4 CladeIB Ancistrorhynchus recurvus 1.079181246 1.380211242 0.4 CladeIB Angraecoides chevalieri 1.079181246 1.204119983 0.4 CladeIB Angraecoides cultriformis 1.477121255 1.342422681 0.4 CladeIB Angraecoides erecta 1.230448921 1.278753601 0.4 CladeIB Angraecoides cf. moandense 1.176091259 1.301029996 0.8 CladeIB Angraecopsis parviflora 1.301029996 1.301029996 0.4 CladeIB Bolusiella iridifolia 0.477121255 0.602059991 0.4 CladeIB Bolusiella maudiae 0.477121255 0.698970004 0.4 CladeIB Calyptrochilum christyanum 1.477121255 1.301029996 0.4 CladeIB Campylocentrum fasciola 0.301029996 0.602059991 0.4 CladeIB Cribbia brachyceras 1.380211242 1.176091259 0.4 CladeIB Cribbia confusa 2.158362492 1.204119983 0.4 CladeIB Cyrtorchis arcuata 2.544068044 1.77815125 0.4 CladeIB Cyrtorchis chailluana 2.431363764 1.954242509 0.4 CladeIB Dendrophylax funalis 2.255272505 1.740362689 0.4 CladeIB Diaphananthe pellucida 2.049218023 1.380211242 0.4 CladeIB Diaphananthe sarcophylla 0.477121255 1.176091259 0.4 CladeIB Diaphananthe vesicata 0.477121255 1.414973348 0.4 CladeIB Dolabrifolia bancoensis 0.477121255 0.301029996 0.4 CladeIB Dolabrifolia cf. aporoides 0 0.84509804 0.4 CladeIB Dolabrifolia disticha 0 0.84509804 0.4 CladeIB Dolabrifolia podochiloides 1.447158031 1.146128036 0.4 CladeIB Eichlerangraecum birrimense 2.908485019 1.278753601 0.4 CladeIB Eichlerangraecum cf. 3.096910013 1.477121255 0.4 eichlerianum CladeIB Eichlerangraecum infundibulare 3.342422681 2.301029996 0.4 CladeIB Eurychone galeandrae 2.954242509 1.397940009 0.4 CladeIB Eurychone rothschildiana 2.880813592 1.361727836 0.4 CladeIB Aerangis ellisii 2.176091259 2.146128036 0.4 CladeIB Aerangis hariotiana 0.77815125 0.602059991 0.4 CladeIB Aerangis hildebrandtii 0.77815125 0.698970004 0.4 CladeIB Aerangis macrocentra 1.77815125 1.812913357 0.4 CladeIB Aerangis punctata 2.176091259 2.079181246 0.4 CladeIB Listrostachys pertusa 1.322219295 0.698970004 0.4 CladeIB Microcoelia gilpinae 0.903089987 1.146128036 0.4 CladeIB Microcoelia stolzii 0.77815125 0.698970004 0.4 CladeIB Mystacidium capense 1.857332496 1.77815125 0.4 CladeIB Mystacidium flanaganii 1.477121255 1.397940009 0.4 CladeIB Pectinariella atlantica 1.602059991 0.602059991 0.4 CladeIB Pectinariella doratophyllum 0.477121255 1 0.4 CladeIB Pectinariella gabonensis 0.77815125 0.477121255 0.4

xxv

CladeIB Pectinariella pungens 1 0.698970004 0.4 CladeIB Pectinariella subulata 0.77815125 0.698970004 0.4 CladeIB Podangis dactyloceras 1.380211242 -1.958607315 0.4 CladeIB Rangaeris amaniensis 1.903089987 2.041392685 0.4 CladeIB Rangaeris muscicola 1.602059991 2.079181246 0.4 CladeIB Rhipidoglossum subsimplex 1.380211242 1.414973348 0.4 CladeIB Rhipidoglossum xanthopollinium 1.447158031 1.397940009 0.4 CladeIB Solenangis clavata 1.431363764 1.431363764 0.4 CladeIB Solenangis wakefieldii 1.301029996 1.740362689 0.4 CladeIB Sphyrarhynchus schliebenii 1.204119983 1.397940009 0.4 CladeIB Tridactyle bicaudata 1.146128036 1.380211242 0.4 CladeIB Tridactyle filifolia 1.301029996 1.361727836 0.4 CladeIB Ypsilopus longifolius 1.301029996 1.698970004 0.4 CladeIB Ypsilopus viridiflorus 1.146128036 1.477121255 0.4 Jumellea Jumellea anjouanensis 1.397940009 1.301029996 0.3 Jumellea Jumellea arachnantha 2.352182518 1.77815125 0.3 Jumellea Jumellea arborescens 2.225309282 2.041392685 0.3 Jumellea Jumellea bosseri 2.336459734 1.477121255 0.3 Jumellea Jumellea brevifolia 2.322219295 2.06069784 0.3 Jumellea Jumellea densifoliata 1.857332496 2 0.3 Jumellea Jumellea exilis 1.748188027 1.792391689 0.3 Jumellea Jumellea francoisii 2 1.397940009 0.3 Jumellea Jumellea hyalina 1.832508913 1.342422681 0.3 Jumellea Jumellea jumelleana 1.880813592 2.041392685 0.3 Jumellea Jumellea longivaginans 1.832508913 1.977723605 0.3 Jumellea Jumellea majalis 2.459392488 2.079181246 0.3 Jumellea Jumellea maxillarioides 2.556302501 1.819543936 0.3 Jumellea Jumellea rigida 2.534026106 2 0.3 Jumellea Jumellea spathulata 1.908485019 1.301029996 0.3 Jumellea Jumellea tenuibracteata 2.352182518 2.041392685 0.3 Jumellea Jumellea teretifolia 2.350248018 2.079181246 0.3 Lemurorchis Lemurorchis madagascariensis 1.255272505 1.176091259 1 Outgroup Acampe ochracea 0.477121255 -1.958607315 1 Outgroup Aerides odorata 2.602059991 1.397940009 1 Outgroup Phalaenopsis cornu-cervi 2.447158031 -1.958607315 1 Outgroup Polystachya fulvilabia 1.447158031 1.477121255 1 Outgroup Vanda tricolor 2.90579588 1 1

xxvi

Annexe 5. ITS2 phylogenetic tree

ITS2 phylogenetic relationships of Angraecinae. 50% Bayesian majority-rule consensus tree from the nuclear ribosomal sequences (ITS2). Values above branches or at nodes represent posterior probability (PP) and bootstrap percentage (BP) support. Dashes represent branches that collapsed in the MP strict consensus tree. Taxa with an asterisk are Angraecum sensu Garay species. Abbreviations in brackets denote sections sensu Garay (1973): Aca = Acaulia, Ang = Angraecum, Arc = Arachnangraecum, Bor = Boryangraecum, Fil = Filangis, Hum = Humblotiangraecum, Pct = Pectinaria, Per = Perrierangraecum, Psj = Pseudojumellea.

xxvii

Angm scottianum (Arc) 0.9 Angm conchiferum (Arc) 51 Angraecum cf. germinyanum (Arc) 0.9/63 clade J Angm germinyanum (Arc) 0.9/76 Angraecum mirabile (Arc)

Angm appendiculatum (Arc)

Aeranthes grandiflora 1/100 Aeranthes arachnites

0.8/62 Angm rutenbergianum (Per) clade D Angm elephantinum (Per)

Sobennikoffia humbertiana

Angm longicalcar (Ang) 1/99 Angm eburneum var. superbum (Ang) clade K Angm eburneum var. eburneum (Ang)

Angraecum chaetopodium (Aca) 1/90 Angraecum ochraceum (Aca) clade M 1/98 Angraecum setipes (Aca)

Angraecum lecomtei (Per) 0.7/83 Angraecum praestans (Hum) 0.7/59 clade D Angraecum drouhardii (Per)

Jumellea rigida 0.7/55 Jumellea francoisii

Lemurorchis madagascariensis

Angm florulentum (Ang)

Oeoniella polystachys

Jumellea teretifolia

Jumellea spathulata

Jumellea majalis

Jumellea longivaginans

Jumellea jumelleana clade II Jumellea hyalina

Jumellea brevifolia

Jumellea bosseri

Jumellea arachnantha

Jumellea densifoliata

Angum dives (Bor) clade H Angm teres (Bor) Angraecum danguyanum (Arc) 1/77 Angraecum panicifolium (Pct) 1/66 clade C 1/98 Angraecum mauritianum (Psj)

Angraecum filicornu (Fil)

Cryptopus paniculatus 1/68 Cryptopus elatus clade A Neobathiea grandidierana

Angraecumpterophyllum*

xxviii

Angraecoides erecta* 1/82 1/95 Angraecoides chevalieri*

Angraecoides cultriformis*

Dendrophylax funalis

Campylocentrumsullivani

Eichlerangraecumeichlerianum*

Aerangis hildebrandtii 0.9/59 Aerangis hariotiana

1/76 Aerangis punctata 1 76 Aerangis ellisii Aerangis macrocentra

Solenangis wakefieldii 1/91 Solenangis clavata

Podangis dactyloceras

Microcoelia stolzii 1/96 Microcoelia gilpinae

1/- ANGRAECINAE Tridactyle filifoli Diaphananthe vesicata 0.9/59 1/82 Diaphananthe sarcophylla

Diaphananthe pellucida

Eurychone rothschildiana clade B Eurychone galeandrae

Calyptrochilumchristyanum

Mystacidium flanaganii 1/100 Mystacidium capense

Cyrtorchis chailluana 0.95/65 Cyrtorchis arcuata

Cribbia confusa 1/100 Cribbia brachyceras

Bolusiella maudiae 1/79 Bolusiella iridifolia

Ypsilopus viridiflorus

Ypsilopus longifolius

Tridactyle bicaudat

Rhipidoglossumxanthopollinium

Rhipidoglossumsubsimplex

Rangaeris muscicola

Rangaeris amaniensis

Neobathiea grandidierana

Listrostachys pertusa

Ancistrorhynchus capitatus

Ancistrorhynchus recurvus

Acampe ochracea outgroup Polystachya fulvilabia

xxix

Annexe 6. cpDNA + morphology phylogenetic tree

Phylogenetic relationships within subtribe Angraecinae. 50% Bayesian majority-rule consensus tree from the combined plastid (matK, rps16 and trnL) and morphological data. Values above branches or at nodes represent posterior probability (PP) and bootstrap percentage (BP) support. Dashes represent branches that collapsed in the MP strict consensus tree. Taxa with an asterisk are Angraecum sensu Garay species. Abbreviations in brackets denote sections sensu Garay (1973): Aca = Acaulia, Agd = Angraecoides, Ang = Angraecum, Arc = Arachnangraecum, Bor = Boryangraecum, Chl = Chlorangraecum, Fil = Filangis, Gom = Gomphocentrum, Had = Hadrangis, Hum = Humblotiangraecum, Lem = Lemurangis, Lep = Lepervenchea, Nan = Nana, Pct = Pectinaria, Per = Perrierangraecum, Psj = Pseudojumellea.

xxx

Angraecum musculiferum tLepp Angraecum caricifolium tLepp Angraecum cf.pauciramosum tLepp Angraecum pauciramosum tLepp 1/100 Angraecum tenuispica tLepp Angraecum madagascariense tLemp Angraecum tenuifolium tLepp Angraecum hermannii tPctp Angraecum baronii tLemp Angraecum costatum tLemp 1/73 Angraecum obversifolium tAgdp Angraecum cf.sacculatum tGomp clade M 1/100 Angraecum pinifolium tBorp Angraecum cf.pinifolium tBorp Angraecum sacciferum tBorp Angraecum cordemoyi tGomp 1/75 Angraecum cf.zeylanicum tGomp 1/92 Angraecum caulescens tGomp 1/93 1 Angraecum huntleyoides tChlp Angraecum calceolus tGomp Angraecum multiflorum tGomp Angraecum ochraceum tAcap 1/93

1/78 Angraecum chaetopodium tAcap Angraecum setipes tAcap 1 Angraecum amplexicaule tPsjp Angraecum serpens tAngp Angraecum penzigianum tAngp 1/90 Angraecum moratii tFilp clade L 1/74 Angraecum florulentum tAngp Angraecum oblongifolium tPerp Angraecum corrugatum tArcp 1/98 1 Angraecum conchoglossum tArcp Angraecum linearifolium tArcp 1/76 Angraecum mirabile tArcp Angraecum conchiferum tArcp 1

1/98 Angraecum expansum tArcp Angraecum germinyanum tArcp clade J 1 Angraecum arachnites tArcp 1/96 Angraecum cf.germinyanum tArcp 1 Angraecum appendiculatum tArcp Angraecum scottianum tArcp 1/74 Angraecum pseudofilicornu tArcp Angraecum longicalcar tAngp Angraecum eburneum var.superbum tAngp 1/98 Angraecum eburneum var.xerophilum tAngp clade K Angraecum eburneum var.eburneum tAngp Oeoniella polystachys Angraecum triangulifolium tAgdp 1 1/100 Angraecum sedifolium tAgdp clade I

1/100 Angraecum pingue tAgdp 1 Angraecum nanum tNanp Angraecum teres tBorp 1/93 clade H Angraecum dives tBorp Angraecum cadetii tHadp 1 1/100 Angraecum bracteosum tHadp clade F Angraecum striatum tHadp Sobennikoffia humbertiana Angraecum.peyrotii.157

xxxi

Angraecum compactum (Per) 1 1/97 Angraecum clareae (Per) 1/61 Angraecum leonis (Hum) clade G 1/100 Angraecum viguieri (Hum) Angraecum magdalenae (Hum) Angraecum sororium (Ang) 1 Angraecum sesquipedale (Ang) clade E Angraecum peyrotii (Per) 1 1 Angraecum cf.peyrotii (Per) Angraecum borbonicum (Per) Angraecum sterophyllum (Per) Angraecum cf.rutenbergianum (Per) Angraecum breve (Per) Angraecum letouzeyi (Per) 1/100 Angraecum cornigerum (Fil) Angraecum liliodorum (Per) 1/97 1/99 Angraecum cf.liliodorum (Per) clade D Angraecum pseudodidieri (Per) Angraecum obesum (Per) Angraecum elephantinum (Per) Angraecum cf.elephantinum (Per) Angraecum didieri (Per) Angraecum ankeranense (Per) 1/- 1/97 Angraecum Angraecum bicallosum (Per) Angraecum drouhardii (Per) 1/99 Angraecum cucullatum (Per) Angraecum lecomtei (Per) Angraecum praestans (Hum) Angraecum panicifolium (Pct) 1 Angraecum cf.humblotianum (Pct) 1/100 Angraecum cf.panicifolium (Pct) Angraecum pectinatum (Pct) Angraecum dasycarpum (Pct) Angraecum ramosum (Arc) clade C 1/75 Angraecum mauritianum (Psj) Angraecum danguyanum (Arc) Angraecum filicornu (Fil) Lemurorchis madagascariensis Jumellea tenuibracteata clade II 1/91 Jumellea rigida 1 Jumellea maxillarioides Jumellea bosseri 1/77 Jumellea arborescens Jumellea francoisii Jumellea exilis Jumellea spathulata 1 Jumellea longivaginans Jumellea Jumellea hyalina 1/97 1 Jumellea densifoliata 1 1/54 Jumellea arachnantha Jumellea jumelleana Jumellea majalis 1/96 Jumellea brevifolia Jumellea anjouanensis Jumellea teretifolia Aeranthes grandiflora 1 Aeranthes arachnites Aeranthes virginalis Aeranthes strangulata 1/100 Aeranthes neoperrieri Aeranthes Aeranthes moratii

1/98 Aeranthes longipes

1/98 Aeranthes caudata Aeranthes aemula Aeranthes tenella Aerangis.punctata

xxxii

Aeranthes.tenella Aerangis punctata

1 Aerangis macrocentra 1/67 Aerangis ellisii 1/87 Aerangis hildebrandtii Aerangis hariotiana Bolusiella maudiae 1/100 Bolusiella iridifolia Eurychone rothschildiana 1/100 Eurychone galeandrae Mystacidium flanaganii 1/100 Mystacidium capense Angraecopsis parviflora ANGRAECINAE Sphyrarhynchus schliebenii 1/100 Ypsilopusviridiflorus 1 Ypsilopuslongifolius Rangaeris amaniensis Rangaeris muscicola Tridactyle filifolia 1/82 Tridactyle bicaudata Listrostachys pertusa 1/100 Cyrtorchis chailluana Cyrtorchis arcuata Pectinariella pungens. 1 Pectinariella gabonensis. Pectinariella atlantica. 1/96 Pectinariella doratophylla. Pectinariella subulata. Dolabrifolia disticha. clade B 1 1 Dolabrifolia bancoensis. 1/100 Dolabrifolia cf.aporoides. Dolabrifolia podochiloides. Solenangis wakefieldii 1/100 Solenangis clavata Microcoelia stolzii 1/100 Microcoelia gilpinae 1/100 Ancistrorhynchus recurvus Ancistrorhynchus capitatus Podangis dactyloceras Rhipidoglossumxanthopollinium 1/100 1 Rhipidoglossumsubsimplex 1/100 Cribbia confusa Cribbia brachyceras Calyptrochilumchristyanum Angraecoides erecta. 1 1/99 Angraecoides cultriformis. Angraecoides chevalieri. 1 1/96 Angraecoides cf.moandense . Dendrophylax funalis 1/96 1/69 Campylocentrumfasciola Eichlerangraecuminfundibulare. 1 1/100 Eichlerangraecumbirrimense. clade I 1/77 Eichlerangraecumcf.eichlerianum. Diaphananthe vesicata 1/82 1/100 Diaphananthe sarcophylla Diaphananthe pellucida Cryptopus paniculatus 1/81 1/80 Cryptopus elatus Oeonia rosea Neobathiea grandidierana Lemurella papillosa 1/99 Lemurella pallidiflora 1 Angraecumrhynchoglossum. (Aca) clade A Angraecumpterophyllum. (Pct) 1/78 1 Angraecum cf.humile. (Lem) Angraecumperparvulum. (Nana) Beclardia macrostachya Erasanthe henricii Vanda tricolor

1/93 Acampe ochracea 1/100 Aerides odorata outgroup Phalaenopsis cornu-cervi Polystachya fulvilabia

xxxiii

Annexe 7. Morphological phylogenetic tree

Phylogenetic relationships of Angraecinae. 50% Bayesian majority-rule consensus tree from morphological data. Values above branches or at nodes represent posterior probability (PP). Dashes represent branches that collapsed in MP strict consensus trees. Taxa with an asterisk are Angraecum sensu Garay species. Abbreviations in brackets denote sections sensu Garay (1973): Aca = Acaulia, Agd = Angraecoides, Ang = Angraecum, Arc = Arachnangraecum, Bor = Boryangraecum, Chl = Chlorangraecum, Fil = Filangis, Gom = Gomphocentrum, Had = Hadrangis, Hum = Humblotiangraecum, Lem = Lemurangis, Lep = Lepervenchea, Pct = Pectinaria, Per = Perrierangraecum, Psj = Pseudojumellea.

xxxiv

Jumellea spathulata Jumellea hyalina Jumellea longivaginans Jumellea exilis Jumellea maxillarioides Jumellea arachnantha Jumellea teretifolia Jumellea tenuibracteata 1 Jumellea rigida Jumellea majalis Jumellea jumelleana Jumellea francoisii Jumellea densifoliata Jumellea brevifolia Jumellea bosseri Jumellea arborescens Jumellea anjouanensis Angraecum viguieri (Hum) 1 Angraecum praestans (Hum) Angraecum magdalenae (Hum) Angraecum compactum (Per) clade G 1 1 Angraecum clareae (Per) Angraecum leonis (Hum) Angraecum longicalcar (Ang) 1 Angraecum eburneum var.xerophilum (Ang) Angraecum eburneum var.superbum (Ang) clade K Angraecum eburneum var.eburneum (Ang) Lemurorchis madagascariensis Angraecum sororium (Ang) clade E Angraecum sesquipedale (Ang) Angraecum germinyanum (Arc) 1 1 Angraecum cf.germinyanum (Arc) Angraecum conchoglossum (Arc) Angraecum arachnites (Arc) Angraecum mirabile (Arc) 0.9 Angraecum conchiferum (Arc) Angraecum expansum (Arc) clade J Angraecum corrugatum (Arc) 1 1 Angraecum appendiculatum (Arc) Angraecum linearifolium (Arc) Angraecum scottianum (Arc) 1 Angraecum pseudofilicornu (Arc) Angraecum oblongifolium (Per) Aeranthes virginalis 1Aeranthes grandiflora Aeranthes caudata Aeranthes longipes Aeranthes arachnites Aeranthes aemula Aeranthes moratii 1 Aeranthes strangulata Aeranthes neoperrieri Aeranthes tenella Microcoelia stolzii 1 Microcoelia gilpinae Campylocentrumsullivanii Dendrophylax funalis Angraecumpterophyllum* 1 Angraecumperparvulum* Angraecum cf.humile* Angraecum striatum (Had) 1 Angraecum cadetii (Had) Angraecum bracteosum (Had) clade F Dolabrifolia disticha* 1Dolabrifolia bancoense* 1 Dolabrifolia cf.aporoides* Dolabrifolia podochiloides* Pectinariella subulata* 1 Pectinariella pungens* Pectinariella gabonense* Pectinariella doratophyllum* Pectinariella atlantica* Angraecum pectinatum (Pct) 1 Angraecum dasycarpum (Pct) Angraecum cf.humblotianum (Pct) Angraecum panicifolium (Pct) 1 clade C 1Angraecum cf.panicifolium (Pct) Angraecum filicornu (Fil) Angraecum danguyanum (Arc) Angraecum musculiferum (Lep) 1 1 Angraecum caricifolium (Lep) Angraecum cf.pauciramosum (Lep) Angraecum pauciramosum (Lep) clade M 1 Angraecum madagascariense (Lem) Angraecum tenuispica (Lep) Angraecum triangulifolium (Agd) 1 Angraecum pingue (Agd) clade I Angraecum sedifolium (Agd) Ypsilopusviridiflorus Angraecumrhynchoglossum* Angraecum caulescens (Gom) 1 Angraecum calceolus (Gom) Angraecum cf.zeylanicum (Gom) clade M Angraecum multiflorum (Gom) Angraecum huntleyoides (Chl) Solenangis wakefieldii 1 Solenangis clavata Rhipidoglossumxanthopollinium 1 ANGRAECINAE Rhipidoglossumsubsimplex Ypsilopuslongifolius Rangaeris muscicola Rangaeris amaniensis Listrostachys pertusa Aerangis hildebrandtii 1 Aerangis hariotiana Diaphananthe vesicata 1 Diaphananthe sarcophylla ngraec es. a

xxxv

Angraecoides erecta* Angraecoides cultriforme* 1 Angraecoides chevalieri* Angraecoides cf.moandense * Cryptopus paniculatus 1 1 Cryptopus elatus Oeonia rosea Ancistrorhynchus recurvus 1 Ancistrorhynchus capitatus Podangis dactyloceras Angraecum drouhardii (Per) 1 Angraecum cucullatum (Per) Angraecum lecomtei (Per) Angraecum didieri (Per) 1 Angraecum ankeranense (Per) Angraecum bicallosum (Per) Angraecum hermannii (Pct) 1 Angraecum baronii (Lem) Angraecum costatum (Lem) Eichlerangraecuminfundibulare* 1 Eichlerangraecumbirrimense* Eichlerangraecumcf.eichlerianum* Tridactyle filifolia 1 Tridactyle bicaudata Lemurella papillosa 1 Lemurella pallidiflora Eurychone rothschildiana 1 Eurychone galeandrae Cyrtorchis chailluana 1 Cyrtorchis arcuata Cribbia confusa 1 Cribbia brachyceras Bolusiella maudiae 1 Bolusiella iridifolia Sobennikoffia humbertiana Beclardia macrostachya Angraecum serpens (Ang) 1 Angraecum penzigianum (Ang) 1 Angraecum pseudodidieri (Per) Angraecum obesum (Per) Angraecum setipes (Aca) 1 Angraecum ochraceum (Aca) Angraecum obversifolium (Agd) Angraecum cf.sacculatum (Gom) Angraecum pinifolium (Bor) 1 Angraecum cf.pinifolium (Bor) Sphyrarhynchus schliebenii Oeoniella polystachys Neobathiea grandidierana Mystacidium flanaganii Mystacidium capense Diaphananthe pellucida Calyptrochilumchristyanum Angraecopsis parviflora Erasanthe henricii Aerangis punctata Aerangis macrocentra Aerangis ellisii Angraecum teres Angraecum tenuifolium (Lep) Angraecum sterophyllum (Per) Angraecum sacciferum (Bor) Angraecum ramosum (Arc) Angraecum nanum (Nana) Angraecum peyrotii (Per) Angraecum moratii (Fil) Angraecum mauritianum (Psj) Angraecum liliodorum (Per) Angraecum letouzeyi (Per) Angraecum florulentum (Ang) Angraecum elephantinum (Per) Angraecum dives (Bor) Angraecum cf.rutenbergianum (Per) Angraecum cf.elephantinum (Per) Angraecum cf.peyrotii (Per) Angraecum cornigerum (Fil) Angraecum cordemoyi (Gom) Angraecum cf.breve (Per) Angraecum cf.liliodorum (Per) Angraecum chaetopodium (Aca) Angraecum borbonicum (Per) Angraecum amplexicaule (Psj) Vanda tricolor 1 Acampe ochracea Aerides odoratum Phalaenopsiscornu-cervi Polystachya fulvilabia

xxxvi

Annexe 8. Dated phylogenetic tree

Maximum credibility tree of the calibrated relaxed molecular clock analysis of Angraecinae inferred from combined plastid matK, rps16 and trnL sequences. Posterior probabilities are displayed above branches in italics; node ages are indicated in bold, with blue bars representing the 95% highest height probability densities (HPD) of the node. Abbreviations: A, Angraecum; At, Aeranthes; J, Jumellea.

xxxvii

Polystachyafulvilabia 3785 27.7 Phalaenopsis cornu9cervi 2 13.4 Vanda tricolor 3785 7.4 3796 9.8 Aerides odorata Acampe ochracea Beclardia macrostachya 3752 15.8 Erasanthe henrici Lemurella papillosa 2 6.9 Lemurella pallidiflora 3796 18.1 3777 Oeonia rosea 13.2 3799 4.4 Cryptopus paniculatus 3799 2.3 3789 9.5 Cryptopus elatus 3789 15 Neobathiea grandidierana 3754 7.6 Bosserangraecum rhynchoglossum Parangraecumperparvulum 3795 12.8 Parangraecumpterophyllum 3799 10.1 Parangraecumcf7humile Diaphananthe pellucida 2 6.8 Diaphananthe vesicata 3799 3.9 Diaphananthe sarcophylla 2 21.1 Dendrophylax funalis 2 8.4 3779 18.5 Campylocentrumfasciola 2 11.8 Angraecoides erecta 2 1.8 Angraecoides cultriformis 2 4.4 2 Angraecoides chevalieri 15.8 2 0.7 Angraecoides cf7moandense Eichlerangraecumcf7eichlerianum 2 0.6Eichlerangreacuminfundibulare 3784 0.4 Eichlerangraecumbirrimense 2 Bolusiella maudiae 34.2 2 8.4 Bolusiella iridifolia 3797 19.5 3736 15.9 Eurychone rothschildiana 2 6.4 Eurychone galeandrae 3785 14.3 Aerangis punctata 2 6.9 Aerangis macrocentra 3778 16 3754 6.7 Aerangis hildebrandtii 2 2.6 3757 6.3 Aerangis hariotiana Aerangis ellisii Dolabrifolia podochiloides 2 4.5 Dolabrifolia cf7aporoides 2 2 2.4 Dolabrifolia disticha 17.6 2 0.9 Dolabrifolia bancoensis Solenangis wakefieldii 2 9.6 Solenangis clavata Microcoelia stolzii 2 9.3 Microcoelia gilpinae 3729 17.4 Rhipidoglossumxanthopollinium 2 7.6 Rhipidoglossumsubsimplex 3767 6.5 2 Cribbia confusa 3725 2.3 17.2 2 11.1 Cribbia brachyceras Mystacidium flanaganii 2 1.8 Mystacidium capense 3796 9 Sphyrarhynchus schliebenii 3786 8.3 Angraecopsis parviflora 3782 16 Calyptrochilumchristyanum 3728 14.6 Ancistrorhynchus recurvus 2 5.6 Ancistrorhynchus capitatus 3797 Podangis dactyloceras 26.1 3752 3725 15.6 10.2 Cyrtorchis chailluana 2 2.3 Cyrtorchis arcuata Rangaeris muscicola 3788 13.3 3744 8.7 Ypsilopus viridiflorus 3767 5.6 3785 6.7 Ypsilopus longifolius Rangaeris amaniensis 3739 13.1 Tridactyle filifolia 2 2.6 Tridactyle bicaudata 3732 13 Listrostachys pertusa 374 11.9 Pectinariellasubulata 2 4.8 Pectinarielladoratophylla 2 2.3 Pectinariellagabonensis 2 1.1 Pectinariellapungens 3749 0.9 Pectinariellaatlantica

35 30 25 20 15 10 50

xxxviii

J.exilis J.tenuibracteata 0.630.39 6.76.1 J.rigida 0.67 4.9 0.21 J.maxillarioides 6.5 J.francoisii 0.69 5 J.bosseri 1 1 2.6 7.8 J.arborescens J.teretifolia J.jumelleana J.longivaginans 0.21 7.5 1 4.8 1 1.5 J.spathulata 0.59 1.1 0.39 4.5 J.hyalina 0.15 7.2 1 J.densifoliata 0.89 1.5 15 J.arachnanta J.majalis 1 2.4 J.brevifolia 0.51 2 J.anjouanensis At.tenella At.moratii At.neoperrieri 0.131 1 10.4 3 2.3 At.strangulata 0.14 2.1 0.2 At.longipes 0.8 1.9 2.3 At.caudata 1 1 17.2 7.5 At.virginalis 1 0.7At.grandiflora 1 0.2 At.arachnites At.aemula Lemurorchis madagascariensis A.filicornu A.ramosum 0.98 5.80.38 A.pectinatum 4.10.72 2.7 A.dasycarpum 0.55 3.2 0.95 4.4 0.76 A.mauritianum 1 1.5 8.8 A.danguyanum 0.82 15.5 A.cf.panicifolium A.panicifolium 1 0.6 A.cf.humblotianum A.sororium 0.89 5.7 A.sesquipedale A.praestans 0.57 14.6 A.lecomtei 0.06 3.1 1 1 A.drouhardii 9.8 3.1 A.cucullatum A.dryadum 1 6.9 A.peyrotii 1 1 A.cf.didieri 1 1 2.9 4.6 A.cf.breve 0.49 0.97 2 14 A.borbonicum

0.28 A.sterophyllum 4.5 0.86 1.4 0.88 A.cf.rutenbergianum 2.4 A.cf.elephantinum 0.96 A.letouzeyi 0.783.83 A.cornigerum

0.21 A.obesum 3.70.38 2.3 A.liliodorum 0.99 1.3 0.88 A.cf.liliodorum 0.13 2.5 13.9 A.elephantinum 0.57 2 A.didieri 0.53 1.5 A.pseudodidieri 0.38 1.3 A.ankeranense A.viguieri 1 5.3 A.magdalenae 0.98 8.5 A.leonis 1 4.7 A.compactum 0.97 0.79 3.9 10.4 A.clareae A.humbertianum 0.64 9.2 A.cadetii 1 1.7 A.striatum 1 0.2 A.bracteosum A.waterlotii.258 35 30 25 20 15 10 50

xxxix

A.nanum

1 0.98 12.2 8.5 A.triangulifolium 0.97 1.6 A.sedifolium 1 2.9 0.2 11.5 A.pingue A.teres 1 6.1 A.dives A.polystachym A.scottianum 0.98 0.84 11.7 5.5 A.pseudofilicornu A.mirabile A.linearifolium 1 0.44 9.3 4.80.57 3.5 A.corrugatum 1 0.4 0.66 4.1 A.conchoglossum 0.98 10.1 A.expansum 1 1 0.8 6.8 A.conchiferum A.bidentantum 1 1.6 A.cf.germinyanum 0.9 1 1.1 4.1 A.arachnites A.appendiculatum 0.08 10.1 A.eburneum var.xerophilum 0.95 1.2 A.eburneum var.longicalcar 0.56 1 0.9 2 A.eburneum var.superbum A.eburneum var.eburneum A.oblongifolium

0.98 6.1 A.penzigianum

0.99 3.9 A.serpens 1 0.8 A.moratii 1 2.4 0.15 9.9 A.florulentum A.multiflorum A.huntleyoides 0.95 6.3 1 2.7 A.cf.zeylanicum 1 0.98 0.8 5.1 A.caulescens A.calceolus A.setipes 0.94 1 9.1 7.3 1 4.7 A.chaetopodium 1 2.8 A.chaetopodium A.costatum 0.85 4.1 A.obversifolium 1 6.2 1 2.1 A.cf.sacculatum A.hermannii 0.24 4.9 A.cordemoyi 0.07 1 5.45.2 A.tenuifolium A.madagascariense 0.84 1 3.4 8.4 A.tenuispica 0.23 0.9 0.93 A.pauciramosum 2.80.19 1.1 0.28 1.2 A.musculiferum 0.14 5.3 1 1.3 A.cf.graminifolium A.caricifolium A.sacciferum 0.15 4.6 A.pinifolium 1 0.28 0.8 4.8 A.cf.pinifolium A.baronii A.amplexicaule

35 30 25 20 15 10 50

Annexe 9. Spur length BAMM analysis

Configuration shifts from the 95% credible set sampled by BAMM for the evolution of spur length across a phylogenetic tree of Angraecinae and evolutionary rates through time. The intensity of colors on branches reflects the instantaneous rate of phenotypic evolution (cool colors = slow, warm = fast). The red curve illustrates the mean speciation rate-through-time trajectory of Angraecinae in million years, and blue that of Vandeae.

xli

Annexe 10. Ancestral state reconstructions of floral traits

Ancestral state reconstructions of floral traits in Angraecinae implemented in ‘diversitree’; colors represent character states and pie charts represent the probability of ancestral states at node. Characters: A, flower color; B, flower size; C, spur length.

xlii

A: color of flower

green white other

xliii

B: flower size

tiny small medium large very large

xliv

C: spur length

very short short medium long very long

xlv

Annexe 11. Speciation analyses

Posterior probability distributions for the speciation rates (in Ma) of morphological characters of subtribe Angraecinae inferred by the MuSSE model. A, leaf surface; B, sepal apex; C, orientation of dorsal sepal; D, orientation of lateral sepal; E, orientation of lateral petal; F, spur apex; G, habitat; H, inflorescence length; I, pedicel length; J, petal apex; K, flower texture; L, spur orientation; M, habit; N, internode; O, stem diameter; P, number of leaves; Q, position of inflorescence; R, leaf apex; S, spur base; T, natural spread (vertical); U, natural spread (horizontal); V, leaf width; W, leaf-length; X, sepal shape; AA, sepal color; AB, petal color; AC, spur color; AD, 3D labellum shape; AE, leaf texture; AF, petal shape; AG, type of inflorescence; AH, spur shape; AI, labellum apex; AJ, 2D labellum shape; AK, leaf shape. Abbreviations: circin, circinnate; coriac, coriaceous; herb, herbaceous; infrafol, infrafoliar; infundi, infundibular; lanceo, lanceolate; memb, membraneous; NA, not applicable; race, raceme; scoop, scoop-shaped; sol, solitary; subter, subterminal; succ, succulent; suprafol, suprafoliar; t, type; tap, tapering; triang, triangular; v, very.

xlvi

xlvii

xlviii

xlix